feat(datasets): periods (start_date/end_date), dg_id column, date filters; rename observation dates

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
2026-09-05 13:16:07 +05:00
parent 37d4f3519b
commit 922812b765
10 changed files with 729 additions and 181 deletions
--- a/README.md
+++ b/README.md
@@ -72,10 +72,10 @@ the KATO column and the year columns:
   `parse_error`).
 3. The client polls `GET /datasets/{id}` until `awaiting_mapping`, then submits
   `POST /datasets/{id}/mapping` with the chosen `kato_column` and a
-   `year_columns` map (each `{column, date}`). Validated against the detected
-   columns; `status` → `extracting`.
+   `year_columns` map (each `{column, start_date}`, optionally `end_date` for
+   a period). Validated against the detected columns; `status` → `extracting`.
 4. A second worker job **unpivots** the attribute table into long-format
-   `dataset_observations` — one row per `(kato_code, date)` with a numeric
+   `dataset_observations` — one row per `(kato_code, start_date)` with a numeric
   `value` (or `value_text` for non-numeric cells); `status` → `ready`. Read
   them via `GET /datasets/{id}/observations` (paginated, optional
   `?kato_code=`).
@@ -84,12 +84,25 @@ the KATO column and the year columns:
 curl -X POST https://dssgis.dwh.kz/datasets/<id>/mapping -H 'Content-Type: application/json' -d '{
  "kato_column": "като",
  "year_columns": [
-    {"column": "F_2023", "date": "2023-01-01"},
-    {"column": "D_2025", "date": "2025-01-01"}
+    {"column": "F_2023", "start_date": "2023-01-01"},
+    {"column": "D_2025", "start_date": "2025-01-01"}
  ]
 }'
 ```

+### Dataset periods
+
+A dataset carries an optional period — `start_date` and `end_date`
+(`YYYY-MM-DD`, both nullable, `end_date` inclusive and never before
+`start_date`) — plus an optional `dg_id` linking it to a *dataset group*: the
+"одноимённые" datasets that describe the same thing for different periods. All
+three are accepted on `POST /datasets` (form fields), `PUT /datasets/{id}` and
+`PATCH /datasets/{id}` (an explicit `null`, or omission on PUT, clears them).
+
+`GET /datasets` filters on them with `?dg_id=`, and with the inclusive bounds
+`?start_date_from=`, `?start_date_to=`, `?end_date_from=` and `?end_date_to=`
+(each `YYYY-MM-DD`).
+
 ## Getting started

 ```sh
@@ -117,13 +130,13 @@ server runs it is served at `/openapi.yaml`, with an interactive **Redoc** UI at
 | GET    | `/categories/{id}`         | get                                  |
 | PUT    | `/categories/{id}`         | update                               |
 | DELETE | `/categories/{id}`         | delete                               |
-| GET    | `/datasets`                | paginated list of summaries (`?page=`, `?page_size=`, `?category_id=`) |
-| POST   | `/datasets`                | upload (multipart: `file`, `file_type`, `category_id`, `code`, `name`, `description?`, `unit?`, `meta?` (JSON), `automated?` (bool)) |
+| GET    | `/datasets`                | paginated list of summaries (`?page=`, `?page_size=`, `?category_id=`, `?dg_id=`, `?start_date_from=`, `?start_date_to=`, `?end_date_from=`, `?end_date_to=`) |
+| POST   | `/datasets`                | upload (multipart: `file`, `file_type`, `category_id`, `code`, `name`, `description?`, `unit?`, `meta?` (JSON), `automated?` (bool), `dg_id?`, `start_date?`, `end_date?`) |
 | GET    | `/datasets/{id}`           | full dataset (geometry as GeoJSON, `bbox` for rasters) |
 | GET    | `/datasets/{id}.geojson`   | GeoJSON `FeatureCollection`; plain `vector` returns its geometry as a single feature with the extracted attribute table as top-level properties; `vector_with_kato` always ignores its own geometry and joins the `districts` table by KATO, mapping observations onto each polygon |
 | GET    | `/datasets/{id}/status`    | processing status; long-polls with `?current=<status>` (holds up to `?wait=` secs, default 25, max 60) |
 | GET    | `/datasets/{id}/download`  | download the stored file             |
-| POST   | `/datasets/{id}/mapping`   | set KATO column + year→date map (vector_with_kato) |
+| POST   | `/datasets/{id}/mapping`   | set KATO column + year→period map (vector_with_kato) |
 | GET    | `/datasets/{id}/observations` | paginated unpivoted values (`?kato_code=`, `?page=`, `?page_size=`) |
 | DELETE | `/datasets/{id}`           | delete (row + object)                |

--- a/api/openapi.yaml
+++ b/api/openapi.yaml
@@ -194,6 +194,41 @@ paths:
            type: string
            enum:
              [pending, parsing, processing, awaiting_mapping, extracting, ready, failed]
+        - name: dg_id
+          in: query
+          required: false
+          description: Filter to a dataset group by id.
+          schema:
+            type: string
+            format: uuid
+        - name: start_date_from
+          in: query
+          required: false
+          description: Keep datasets whose `start_date` is on or after this date.
+          schema:
+            type: string
+            format: date
+        - name: start_date_to
+          in: query
+          required: false
+          description: Keep datasets whose `start_date` is on or before this date.
+          schema:
+            type: string
+            format: date
+        - name: end_date_from
+          in: query
+          required: false
+          description: Keep datasets whose `end_date` is on or after this date.
+          schema:
+            type: string
+            format: date
+        - name: end_date_to
+          in: query
+          required: false
+          description: Keep datasets whose `end_date` is on or before this date.
+          schema:
+            type: string
+            format: date
      responses:
        "200":
          description: A page of dataset summaries
@@ -240,6 +275,20 @@ paths:
                  description: Arbitrary user JSON (sent as a JSON string).
                automated:
                  type: boolean
+                dg_id:
+                  type: string
+                  format: uuid
+                  description: Optional dataset group the dataset belongs to.
+                start_date:
+                  type: string
+                  format: date
+                  description: Start of the period the dataset describes.
+                end_date:
+                  type: string
+                  format: date
+                  description: >-
+                    End of the period (inclusive); must not be before
+                    `start_date`.
      responses:
        "201":
          description: Dataset created; processing enqueued
@@ -275,7 +324,8 @@ paths:
        Updates category, name, description, and unit. Every field is
        required; the file, file type, and processing/lifecycle state are
        never touched here. Sending an empty string for `description` or
-        `unit` clears that (optional) field.
+        `unit` clears that (optional) field. `dg_id`, `start_date` and
+        `end_date` are optional: omitting one clears that (nullable) column.
      requestBody:
        required: true
        content:
@@ -295,9 +345,9 @@ paths:
      summary: Partially update a dataset's editable metadata
      description: |
        Updates only the fields present in the body, from category, name,
-        description, and unit. The file, file type, and processing/lifecycle
-        state are never touched here. Sending an explicit `null` for
-        `description` or `unit` clears that (optional) field; omitting a
+        description, unit, `dg_id`, `start_date` and `end_date`. The file,
+        file type, and processing/lifecycle state are never touched here.
+        Sending an explicit `null` for a nullable field clears it; omitting a
        field leaves it unchanged.
      requestBody:
        required: true
@@ -492,9 +542,9 @@ paths:
        the dataset's detected `attribute_columns`. Moves the dataset to
        `extracting`; observations are produced asynchronously.

-        Each year column maps to either a single point in time (`date` only,
-        the original behavior) or a period (`date` and `date_to`, inclusive).
-        `date_to`, when present, must not be before `date`.
+        Each year column maps to either a single point in time (`start_date`
+        only, the original behavior) or a period (`start_date` and `end_date`,
+        inclusive). `end_date`, when present, must not be before `start_date`.
      requestBody:
        required: true
        content:
@@ -685,6 +735,20 @@ components:
        unit:
          type: string
          description: An empty string clears the field.
+        dg_id:
+          type: [string, "null"]
+          format: uuid
+          description: Dataset group; omitted or null clears the field.
+        start_date:
+          type: [string, "null"]
+          format: date
+          description: Start of the dataset's period; omitted or null clears it.
+        end_date:
+          type: [string, "null"]
+          format: date
+          description: >-
+            End of the dataset's period (inclusive); omitted or null clears it.
+            Must not be before `start_date`.

    DatasetPatch:
      type: object
@@ -703,6 +767,20 @@ components:
        unit:
          type: [string, "null"]
          description: null clears the field.
+        dg_id:
+          type: [string, "null"]
+          format: uuid
+          description: null clears the field.
+        start_date:
+          type: [string, "null"]
+          format: date
+          description: null clears the field.
+        end_date:
+          type: [string, "null"]
+          format: date
+          description: >-
+            null clears the field. Must not be before the resulting
+            `start_date`.

    AttributeColumn:
      type: object
@@ -717,23 +795,23 @@ components:

    YearColumn:
      type: object
-      required: [column, date]
+      required: [column, start_date]
      properties:
        column:
          type: string
-        date:
+        start_date:
          type: string
          format: date
          description: >-
-            Point in time, or the start of the period when `date_to` is also
+            Point in time, or the start of the period when `end_date` is also
            set.
-        date_to:
+        end_date:
          type: [string, "null"]
          format: date
          description: >-
-            Optional period end (inclusive), must not be before `date`. Omitted
-            or null means a single point in time (the original, pre-period
-            behavior).
+            Optional period end (inclusive), must not be before `start_date`.
+            Omitted or null means a single point in time (the original,
+            pre-period behavior).

    GeoJSONGeometry:
      type: [object, "null"]
@@ -763,7 +841,7 @@ components:
            mapping to its date->value pairs.

            For a `vector_with_kato` dataset mapped with one or more period
-            year columns (`date_to` set), a sibling `periods` object is also
+            year columns (`end_date` set), a sibling `periods` object is also
            present, mapping each period's start date (the same key used in
            `data`) to its end date. `periods` is omitted entirely when the
            dataset has no period mappings.
@@ -792,6 +870,19 @@ components:
        category_id:
          type: string
          format: uuid
+        dg_id:
+          type: [string, "null"]
+          format: uuid
+          description: The dataset group this dataset belongs to, if any.
+        start_date:
+          type: [string, "null"]
+          format: date
+          description: Start of the period the dataset describes.
+        end_date:
+          type: [string, "null"]
+          format: date
+          description: >-
+            End of the period (inclusive). Drives the "freshest in group" rule.
        name:
          type: string
        description:
@@ -824,6 +915,19 @@ components:
        category_id:
          type: string
          format: uuid
+        dg_id:
+          type: [string, "null"]
+          format: uuid
+          description: The dataset group this dataset belongs to, if any.
+        start_date:
+          type: [string, "null"]
+          format: date
+          description: Start of the period the dataset describes.
+        end_date:
+          type: [string, "null"]
+          format: date
+          description: >-
+            End of the period (inclusive). Drives the "freshest in group" rule.
        name:
          type: string
        description:
@@ -901,12 +1005,12 @@ components:
      examples:
        - kato_column: "като"
          year_columns:
-            - { column: F_2023, date: "2023-01-01" }
-            - { column: D_2025, date: "2025-01-01" }
+            - { column: F_2023, start_date: "2023-01-01" }
+            - { column: D_2025, start_date: "2025-01-01" }

    Observation:
      type: object
-      required: [id, dataset_id, kato_code, date]
+      required: [id, dataset_id, kato_code, start_date]
      properties:
        id:
          type: string
@@ -916,15 +1020,15 @@ components:
          format: uuid
        kato_code:
          type: string
-        date:
+        start_date:
          type: string
          format: date
-        date_to:
+        end_date:
          type: [string, "null"]
          format: date
          description: >-
            Period end (inclusive), carried over from the year mapping's
-            `date_to`. Null for a point-in-time observation.
+            `end_date`. Null for a point-in-time observation.
        value:
          type: [number, "null"]
          description: Numeric cell value, or null when non-numeric/empty.
--- a/internal/domain/dataset.go
+++ b/internal/domain/dataset.go
@@ -71,20 +71,28 @@ type DatasetFilter struct {
 	FileType   *FileType
 	Automated  *bool
 	Status     *string
+	// DatasetGroupID restricts the listing to one dataset group.
+	DatasetGroupID *uuid.UUID
+	// StartDateFrom / StartDateTo and EndDateFrom / EndDateTo are inclusive
+	// YYYY-MM-DD bounds on the dataset's period columns.
+	StartDateFrom *string
+	StartDateTo   *string
+	EndDateFrom   *string
+	EndDateTo     *string
 }

 // Observation is a single unpivoted value from a dataset's attribute table,
-// keyed by KATO code and date (optionally a date..DateTo period). Exactly one
-// of Value / ValueText is typically set (numeric vs non-numeric cell); both
+// keyed by KATO code and date (optionally a StartDate..EndDate period). Exactly
+// one of Value / ValueText is typically set (numeric vs non-numeric cell); both
 // may be nil for an empty cell.
 type Observation struct {
 	ID        uuid.UUID `json:"id"`
 	DatasetID uuid.UUID `json:"dataset_id"`
 	KatoCode  string    `json:"kato_code"`
-	Date      string    `json:"date"`
-	// DateTo is the period's end date. Nil for a point-in-time observation
+	StartDate string    `json:"start_date"`
+	// EndDate is the period's end date. Nil for a point-in-time observation
 	// (pre-existing behavior).
-	DateTo    *string  `json:"date_to,omitempty"`
+	EndDate   *string  `json:"end_date,omitempty"`
 	Value     *float64 `json:"value"`
 	ValueText *string  `json:"value_text"`
 }
@@ -185,21 +193,29 @@ type AttributeColumn struct {
 }

 // YearColumn maps an attribute column to the date (or date period) it
-// represents, e.g. {"column": "F_2023", "date": "2023-01-01"} for a single
-// point in time, or {"column": "F_2020_2023", "date": "2020-01-01", "date_to":
-// "2023-12-31"} for a period. DateTo is optional; when nil the mapping is a
-// single point in time, preserving the original (pre-period) behavior.
+// represents, e.g. {"column": "F_2023", "start_date": "2023-01-01"} for a
+// single point in time, or {"column": "F_2020_2023", "start_date":
+// "2020-01-01", "end_date": "2023-12-31"} for a period. EndDate is optional;
+// when nil the mapping is a single point in time, preserving the original
+// (pre-period) behavior.
 type YearColumn struct {
-	Column string  `json:"column"`
-	Date   string  `json:"date"`
-	DateTo *string `json:"date_to,omitempty"`
+	Column    string  `json:"column"`
+	StartDate string  `json:"start_date"`
+	EndDate   *string `json:"end_date,omitempty"`
 }

 // DatasetSummary is the lightweight view of a dataset used in list responses.
 // It omits the heavy geometry/attribute/JSONB fields.
 type DatasetSummary struct {
-	ID          uuid.UUID `json:"id"`
-	CategoryID  uuid.UUID `json:"category_id"`
+	ID         uuid.UUID `json:"id"`
+	CategoryID uuid.UUID `json:"category_id"`
+	// DatasetGroupID links the dataset to its dataset group ("одноимённые"
+	// datasets versioned by period). Nullable.
+	DatasetGroupID *uuid.UUID `json:"dg_id"`
+	// StartDate / EndDate is the period the dataset describes (YYYY-MM-DD).
+	// Both nullable; EndDate drives the "freshest in group" rule.
+	StartDate   *string   `json:"start_date"`
+	EndDate     *string   `json:"end_date"`
 	Name        string    `json:"name"`
 	Description *string   `json:"description"`
 	Unit        *string   `json:"unit"`
@@ -212,13 +228,20 @@ type DatasetSummary struct {

 // Dataset is a geo file stored in the object store and grouped under a category.
 type Dataset struct {
-	ID          uuid.UUID `json:"id"`
-	CategoryID  uuid.UUID `json:"category_id"`
-	Name        string    `json:"name"`
-	Description *string   `json:"description"`
-	Unit        *string   `json:"unit"`
-	Filename    string    `json:"filename"`
-	StorageKey  string    `json:"storage_key"`
+	ID         uuid.UUID `json:"id"`
+	CategoryID uuid.UUID `json:"category_id"`
+	// DatasetGroupID links the dataset to its dataset group ("одноимённые"
+	// datasets versioned by period). Nullable.
+	DatasetGroupID *uuid.UUID `json:"dg_id"`
+	// StartDate / EndDate is the period the dataset describes (YYYY-MM-DD).
+	// Both nullable; EndDate drives the "freshest in group" rule.
+	StartDate   *string `json:"start_date"`
+	EndDate     *string `json:"end_date"`
+	Name        string  `json:"name"`
+	Description *string `json:"description"`
+	Unit        *string `json:"unit"`
+	Filename    string  `json:"filename"`
+	StorageKey  string  `json:"storage_key"`
 	// CogStorageKey points to the Cloud-Optimized GeoTIFF for rasters. Nullable.
 	CogStorageKey *string  `json:"cog_storage_key"`
 	FileType      FileType `json:"file_type"`
@@ -262,3 +285,37 @@ type Dataset struct {
 	CreatedAt time.Time `json:"created_at"`
 	UpdatedAt time.Time `json:"updated_at"`
 }
+
+// DatasetFields is the set of editable metadata columns written by
+// UpdateFields. Nil pointers clear the corresponding nullable column.
+type DatasetFields struct {
+	CategoryID     uuid.UUID
+	Name           string
+	Description    *string
+	Unit           *string
+	DatasetGroupID *uuid.UUID
+	StartDate      *string
+	EndDate        *string
+}
+
+// ValidateDatePeriod checks that start/end (YYYY-MM-DD, either may be nil)
+// parse and that end is not before start. Returns an ErrValidation-wrapped
+// error naming the offending field.
+func ValidateDatePeriod(start, end *string) error {
+	var from, to time.Time
+	var err error
+	if start != nil {
+		if from, err = time.Parse("2006-01-02", *start); err != nil {
+			return fmt.Errorf("%w: start_date %q is not YYYY-MM-DD", ErrValidation, *start)
+		}
+	}
+	if end != nil {
+		if to, err = time.Parse("2006-01-02", *end); err != nil {
+			return fmt.Errorf("%w: end_date %q is not YYYY-MM-DD", ErrValidation, *end)
+		}
+	}
+	if start != nil && end != nil && to.Before(from) {
+		return fmt.Errorf("%w: end_date %q is before start_date %q", ErrValidation, *end, *start)
+	}
+	return nil
+}
--- a/internal/domain/dataset_test.go
+++ b/internal/domain/dataset_test.go
@@ -1,6 +1,9 @@
 package domain

-import "testing"
+import (
+	"errors"
+	"testing"
+)

 func TestFileTypeValid(t *testing.T) {
 	valid := []FileType{FileTypeVectorWithKato, FileTypeVector, FileTypeRaster}
@@ -72,3 +75,31 @@ func TestExtensionAllowedFor(t *testing.T) {
 		}
 	}
 }
+
+func TestValidateDatePeriod(t *testing.T) {
+	s := func(v string) *string { return &v }
+	cases := []struct {
+		name       string
+		start, end *string
+		wantErr    bool
+	}{
+		{"both nil", nil, nil, false},
+		{"start only", s("2020-01-01"), nil, false},
+		{"ordered", s("2020-01-01"), s("2020-12-31"), false},
+		{"same day", s("2020-01-01"), s("2020-01-01"), false},
+		{"reversed", s("2020-12-31"), s("2020-01-01"), true},
+		{"bad start", s("01.01.2020"), nil, true},
+		{"bad end", nil, s("2020-13-01"), true},
+	}
+	for _, tc := range cases {
+		t.Run(tc.name, func(t *testing.T) {
+			err := ValidateDatePeriod(tc.start, tc.end)
+			if (err != nil) != tc.wantErr {
+				t.Fatalf("err=%v wantErr=%v", err, tc.wantErr)
+			}
+			if err != nil && !errors.Is(err, ErrValidation) {
+				t.Fatalf("expected ErrValidation, got %v", err)
+			}
+		})
+	}
+}
--- a/internal/repository/postgres/dataset.go
+++ b/internal/repository/postgres/dataset.go
@@ -27,7 +27,7 @@ func NewDatasetRepository(pool *pgxpool.Pool) *DatasetRepository {
 // datasetColumns lists the dataset columns for SELECT and RETURNING. The
 // geometry is exposed as GeoJSON (jsonb) rather than its raw EWKB form, and a
 // bounding box array is derived for raster datasets only.
-const datasetColumns = `id, category_id, name, description, unit, filename, storage_key, cog_storage_key, file_type, size_bytes, content_type, properties, meta, automated, status, attribute_columns, kato_column, year_columns, parse_error, ST_AsGeoJSON(geometry)::jsonb AS geometry,
+const datasetColumns = `id, category_id, dg_id, to_char(start_date, 'YYYY-MM-DD') AS start_date, to_char(end_date, 'YYYY-MM-DD') AS end_date, name, description, unit, filename, storage_key, cog_storage_key, file_type, size_bytes, content_type, properties, meta, automated, status, attribute_columns, kato_column, year_columns, parse_error, ST_AsGeoJSON(geometry)::jsonb AS geometry,
 	CASE WHEN file_type = 'raster' AND geometry IS NOT NULL
 	     THEN ARRAY[ST_XMin(geometry), ST_YMin(geometry), ST_XMax(geometry), ST_YMax(geometry)]
 	     ELSE NULL END AS bbox,
@@ -37,7 +37,7 @@ const datasetColumns = `id, category_id, name, description, unit, filename, stor
 func scanDataset(row pgx.Row) (domain.Dataset, error) {
 	var d domain.Dataset
 	err := row.Scan(
-		&d.ID, &d.CategoryID, &d.Name, &d.Description, &d.Unit,
+		&d.ID, &d.CategoryID, &d.DatasetGroupID, &d.StartDate, &d.EndDate, &d.Name, &d.Description, &d.Unit,
 		&d.Filename, &d.StorageKey, &d.CogStorageKey, &d.FileType, &d.SizeBytes, &d.ContentType,
 		&d.Properties, &d.Meta, &d.Automated, &d.Status,
 		&d.AttributeColumns, &d.KatoColumn, &d.YearColumns, &d.ParseError,
@@ -59,28 +59,28 @@ func nullableJSON(raw json.RawMessage) any {
 // Create inserts a new dataset and returns the stored row.
 func (r *DatasetRepository) Create(ctx context.Context, d domain.Dataset) (domain.Dataset, error) {
 	row := r.pool.QueryRow(ctx,
-		`INSERT INTO datasets (category_id, name, description, unit, filename, storage_key, file_type, size_bytes, content_type, properties, meta, automated, status)
-		 VALUES ($1, $2, $3, $4, $5, $6, $7, $8, $9, $10, $11, $12, $13)
+		`INSERT INTO datasets (category_id, name, description, unit, filename, storage_key, file_type, size_bytes, content_type, properties, meta, automated, status, dg_id, start_date, end_date)
+		 VALUES ($1, $2, $3, $4, $5, $6, $7, $8, $9, $10, $11, $12, $13, $14, $15::date, $16::date)
 		 RETURNING `+datasetColumns,
 		d.CategoryID, d.Name, d.Description, d.Unit, d.Filename, d.StorageKey, d.FileType, d.SizeBytes, d.ContentType,
 		nullableJSON(d.Properties), nullableJSON(d.Meta), d.Automated, d.Status,
+		d.DatasetGroupID, d.StartDate, d.EndDate,
 	)
 	out, err := scanDataset(row)
 	return out, mapError(err)
 }

-// UpdateFields overwrites a dataset's editable metadata columns (category,
-// name, description, unit). description and unit are nullable: a nil pointer
-// clears the column. Returns domain.ErrNotFound if the dataset does not
-// exist, or domain.ErrConflict if categoryID does not reference an existing
-// category.
-func (r *DatasetRepository) UpdateFields(ctx context.Context, id, categoryID uuid.UUID, name string, description, unit *string) (domain.Dataset, error) {
+// UpdateFields overwrites a dataset's editable metadata columns. Nil pointers
+// clear the nullable columns. Returns domain.ErrNotFound if the dataset does
+// not exist, or domain.ErrConflict if the category/group does not exist.
+func (r *DatasetRepository) UpdateFields(ctx context.Context, id uuid.UUID, f domain.DatasetFields) (domain.Dataset, error) {
 	row := r.pool.QueryRow(ctx,
 		`UPDATE datasets
-		 SET category_id = $2, name = $3, description = $4, unit = $5, updated_at = now()
+		 SET category_id = $2, name = $3, description = $4, unit = $5,
+		     dg_id = $6, start_date = $7::date, end_date = $8::date, updated_at = now()
 		 WHERE id = $1
 		 RETURNING `+datasetColumns,
-		id, categoryID, name, description, unit,
+		id, f.CategoryID, f.Name, f.Description, f.Unit, f.DatasetGroupID, f.StartDate, f.EndDate,
 	)
 	out, err := scanDataset(row)
 	return out, mapError(err)
@@ -276,23 +276,23 @@ func (r *DatasetRepository) ReplaceObservations(ctx context.Context, datasetID u
 	if len(obs) > 0 {
 		rows := make([][]any, len(obs))
 		for i, o := range obs {
-			d, err := time.Parse("2006-01-02", o.Date)
+			d, err := time.Parse("2006-01-02", o.StartDate)
 			if err != nil {
-				return fmt.Errorf("invalid observation date %q: %w", o.Date, err)
+				return fmt.Errorf("invalid observation start_date %q: %w", o.StartDate, err)
 			}
-			var dateTo any
-			if o.DateTo != nil {
-				dt, err := time.Parse("2006-01-02", *o.DateTo)
+			var endDate any
+			if o.EndDate != nil {
+				dt, err := time.Parse("2006-01-02", *o.EndDate)
 				if err != nil {
-					return fmt.Errorf("invalid observation date_to %q: %w", *o.DateTo, err)
+					return fmt.Errorf("invalid observation end_date %q: %w", *o.EndDate, err)
 				}
-				dateTo = dt
+				endDate = dt
 			}
-			rows[i] = []any{datasetID, o.KatoCode, d, dateTo, o.Value, o.ValueText}
+			rows[i] = []any{datasetID, o.KatoCode, d, endDate, o.Value, o.ValueText}
 		}
 		_, err := tx.CopyFrom(ctx,
 			pgx.Identifier{"dataset_observations"},
-			[]string{"dataset_id", "kato_code", "date", "date_to", "value", "value_text"},
+			[]string{"dataset_id", "kato_code", "start_date", "end_date", "value", "value_text"},
 			pgx.CopyFromRows(rows),
 		)
 		if err != nil {
@@ -303,16 +303,16 @@ func (r *DatasetRepository) ReplaceObservations(ctx context.Context, datasetID u
 	return mapError(tx.Commit(ctx))
 }

-const observationColumns = `id, dataset_id, kato_code, to_char(date, 'YYYY-MM-DD') AS date, to_char(date_to, 'YYYY-MM-DD') AS date_to, value, value_text`
+const observationColumns = `id, dataset_id, kato_code, to_char(start_date, 'YYYY-MM-DD') AS start_date, to_char(end_date, 'YYYY-MM-DD') AS end_date, value, value_text`

 func scanObservation(row pgx.Row) (domain.Observation, error) {
 	var o domain.Observation
-	err := row.Scan(&o.ID, &o.DatasetID, &o.KatoCode, &o.Date, &o.DateTo, &o.Value, &o.ValueText)
+	err := row.Scan(&o.ID, &o.DatasetID, &o.KatoCode, &o.StartDate, &o.EndDate, &o.Value, &o.ValueText)
 	return o, err
 }

 // ListObservations returns a page of observations for a dataset, optionally
-// filtered by KATO code, ordered by (kato_code, date).
+// filtered by KATO code, ordered by (kato_code, start_date).
 func (r *DatasetRepository) ListObservations(ctx context.Context, datasetID uuid.UUID, katoCode *string, limit, offset int) ([]domain.Observation, error) {
 	base := `SELECT ` + observationColumns + ` FROM dataset_observations WHERE dataset_id = $1`

@@ -322,11 +322,11 @@ func (r *DatasetRepository) ListObservations(ctx context.Context, datasetID uuid
 	)
 	if katoCode != nil {
 		rows, err = r.pool.Query(ctx,
-			base+` AND kato_code = $2 ORDER BY kato_code, date LIMIT $3 OFFSET $4`,
+			base+` AND kato_code = $2 ORDER BY kato_code, start_date LIMIT $3 OFFSET $4`,
 			datasetID, *katoCode, limit, offset)
 	} else {
 		rows, err = r.pool.Query(ctx,
-			base+` ORDER BY kato_code, date LIMIT $2 OFFSET $3`,
+			base+` ORDER BY kato_code, start_date LIMIT $2 OFFSET $3`,
 			datasetID, limit, offset)
 	}
 	if err != nil {
@@ -368,12 +368,25 @@ func (r *DatasetRepository) GetByID(ctx context.Context, id uuid.UUID) (domain.D
 	return out, mapError(err)
 }

-const datasetSummaryColumns = `id, category_id, name, description, unit, file_type, size_bytes, status, created_at, updated_at`
+// LatestInGroup returns the freshest dataset of a group: greatest end_date
+// (NULL end_date sorts last), ties broken by created_at DESC. status and
+// fileType optionally restrict candidates. Returns domain.ErrNotFound when no
+// dataset matches.
+func (r *DatasetRepository) LatestInGroup(ctx context.Context, groupID uuid.UUID, status *string, fileType *domain.FileType) (domain.Dataset, error) {
+	where, args := datasetFilterClause(domain.DatasetFilter{DatasetGroupID: &groupID, Status: status, FileType: fileType}, nil)
+	row := r.pool.QueryRow(ctx,
+		`SELECT `+datasetColumns+` FROM datasets`+where+
+			` ORDER BY end_date DESC NULLS LAST, created_at DESC LIMIT 1`, args...)
+	out, err := scanDataset(row)
+	return out, mapError(err)
+}
+
+const datasetSummaryColumns = `id, category_id, dg_id, to_char(start_date, 'YYYY-MM-DD') AS start_date, to_char(end_date, 'YYYY-MM-DD') AS end_date, name, description, unit, file_type, size_bytes, status, created_at, updated_at`

 func scanDatasetSummary(row pgx.Row) (domain.DatasetSummary, error) {
 	var d domain.DatasetSummary
 	err := row.Scan(
-		&d.ID, &d.CategoryID, &d.Name, &d.Description, &d.Unit,
+		&d.ID, &d.CategoryID, &d.DatasetGroupID, &d.StartDate, &d.EndDate, &d.Name, &d.Description, &d.Unit,
 		&d.FileType, &d.SizeBytes, &d.Status, &d.CreatedAt, &d.UpdatedAt,
 	)
 	return d, err
@@ -400,6 +413,22 @@ func datasetFilterClause(f domain.DatasetFilter, args []any) (string, []any) {
 		args = append(args, *f.Status)
 		conds = append(conds, fmt.Sprintf("status = $%d", len(args)))
 	}
+	if f.DatasetGroupID != nil {
+		args = append(args, *f.DatasetGroupID)
+		conds = append(conds, fmt.Sprintf("dg_id = $%d", len(args)))
+	}
+	for _, b := range []struct {
+		col, op string
+		val     *string
+	}{
+		{"start_date", ">=", f.StartDateFrom}, {"start_date", "<=", f.StartDateTo},
+		{"end_date", ">=", f.EndDateFrom}, {"end_date", "<=", f.EndDateTo},
+	} {
+		if b.val != nil {
+			args = append(args, *b.val)
+			conds = append(conds, fmt.Sprintf("%s %s $%d::date", b.col, b.op, len(args)))
+		}
+	}
 	if len(conds) == 0 {
 		return "", args
 	}
--- a/internal/repository/postgres/district.go
+++ b/internal/repository/postgres/district.go
@@ -9,14 +9,14 @@ import (
 )

 // ListAllObservations returns every observation for a dataset ordered by
-// (kato_code, date). Unlike ListObservations it is unpaginated: it backs GeoJSON
+// (kato_code, start_date). Unlike ListObservations it is unpaginated: it backs GeoJSON
 // assembly, where all of a dataset's observations are needed at once.
 func (r *DatasetRepository) ListAllObservations(ctx context.Context, datasetID uuid.UUID) ([]domain.Observation, error) {
 	rows, err := r.pool.Query(ctx,
 		`SELECT `+observationColumns+`
 		 FROM dataset_observations
 		 WHERE dataset_id = $1
-		 ORDER BY kato_code, date`,
+		 ORDER BY kato_code, start_date`,
 		datasetID)
 	if err != nil {
 		return nil, mapError(err)
--- a/internal/service/dataset.go
+++ b/internal/service/dataset.go
@@ -37,7 +37,8 @@ var (
 type DatasetRepository interface {
 	Create(ctx context.Context, d domain.Dataset) (domain.Dataset, error)
 	GetByID(ctx context.Context, id uuid.UUID) (domain.Dataset, error)
-	UpdateFields(ctx context.Context, id, categoryID uuid.UUID, name string, description, unit *string) (domain.Dataset, error)
+	UpdateFields(ctx context.Context, id uuid.UUID, f domain.DatasetFields) (domain.Dataset, error)
+	LatestInGroup(ctx context.Context, groupID uuid.UUID, status *string, fileType *domain.FileType) (domain.Dataset, error)
 	ListSummaries(ctx context.Context, filter domain.DatasetFilter, limit, offset int) ([]domain.DatasetSummary, error)
 	Count(ctx context.Context, filter domain.DatasetFilter) (int, error)
 	Delete(ctx context.Context, id uuid.UUID) error
@@ -125,6 +126,11 @@ type UploadInput struct {
 	ContentType string
 	Size        int64
 	Reader      io.Reader
+	// DatasetGroupID optionally files the dataset under a dataset group.
+	DatasetGroupID *uuid.UUID
+	// StartDate / EndDate is the period the dataset describes (YYYY-MM-DD).
+	StartDate *string
+	EndDate   *string
 }

 // DatasetService implements dataset business rules and object storage handling.
@@ -190,6 +196,9 @@ func (s *DatasetService) Upload(ctx context.Context, in UploadInput) (domain.Dat
 	if err := s.ensureCategoryExists(ctx, in.CategoryID); err != nil {
 		return domain.Dataset{}, err
 	}
+	if err := domain.ValidateDatePeriod(in.StartDate, in.EndDate); err != nil {
+		return domain.Dataset{}, err
+	}

 	name := in.Name
 	if name == "" {
@@ -210,18 +219,21 @@ func (s *DatasetService) Upload(ctx context.Context, in UploadInput) (domain.Dat
 	}

 	dataset, err := s.repo.Create(ctx, domain.Dataset{
-		CategoryID:  in.CategoryID,
-		Name:        name,
-		Description: in.Description,
-		Unit:        in.Unit,
-		Meta:        in.Meta,
-		Automated:   in.Automated,
-		Status:      status,
-		Filename:    in.Filename,
-		StorageKey:  storageKey,
-		FileType:    in.FileType,
-		SizeBytes:   in.Size,
-		ContentType: in.ContentType,
+		CategoryID:     in.CategoryID,
+		DatasetGroupID: in.DatasetGroupID,
+		StartDate:      in.StartDate,
+		EndDate:        in.EndDate,
+		Name:           name,
+		Description:    in.Description,
+		Unit:           in.Unit,
+		Meta:           in.Meta,
+		Automated:      in.Automated,
+		Status:         status,
+		Filename:       in.Filename,
+		StorageKey:     storageKey,
+		FileType:       in.FileType,
+		SizeBytes:      in.Size,
+		ContentType:    in.ContentType,
 	})
 	if err != nil {
 		// Compensate: the row was not written, so the object would be orphaned.
@@ -682,17 +694,17 @@ func (s *DatasetService) SaveMapping(ctx context.Context, id uuid.UUID, in Mappi
 		if _, ok := known[yc.Column]; !ok {
 			return domain.Dataset{}, fmt.Errorf("%w: year column %q is not among the detected columns", domain.ErrValidation, yc.Column)
 		}
-		from, err := time.Parse("2006-01-02", yc.Date)
+		from, err := time.Parse("2006-01-02", yc.StartDate)
 		if err != nil {
-			return domain.Dataset{}, fmt.Errorf("%w: invalid date %q for column %q (want YYYY-MM-DD)", domain.ErrValidation, yc.Date, yc.Column)
+			return domain.Dataset{}, fmt.Errorf("%w: invalid start_date %q for column %q (want YYYY-MM-DD)", domain.ErrValidation, yc.StartDate, yc.Column)
 		}
-		if yc.DateTo != nil {
-			to, err := time.Parse("2006-01-02", *yc.DateTo)
+		if yc.EndDate != nil {
+			to, err := time.Parse("2006-01-02", *yc.EndDate)
 			if err != nil {
-				return domain.Dataset{}, fmt.Errorf("%w: invalid date_to %q for column %q (want YYYY-MM-DD)", domain.ErrValidation, *yc.DateTo, yc.Column)
+				return domain.Dataset{}, fmt.Errorf("%w: invalid end_date %q for column %q (want YYYY-MM-DD)", domain.ErrValidation, *yc.EndDate, yc.Column)
 			}
 			if to.Before(from) {
-				return domain.Dataset{}, fmt.Errorf("%w: date_to %q is before date %q for column %q", domain.ErrValidation, *yc.DateTo, yc.Date, yc.Column)
+				return domain.Dataset{}, fmt.Errorf("%w: end_date %q is before start_date %q for column %q", domain.ErrValidation, *yc.EndDate, yc.StartDate, yc.Column)
 			}
 		}
 	}
@@ -748,9 +760,10 @@ func (s *DatasetService) Extract(ctx context.Context, id uuid.UUID) error {
 }

 // buildObservations unpivots rows into observations. Rows without a KATO code
-// are skipped; duplicate (kato, date, date_to) tuples keep the last value.
-// Numeric cells populate Value, others ValueText. A year column's DateTo (nil
-// for a point-in-time mapping) carries through to the observation unchanged.
+// are skipped; duplicate (kato, start_date, end_date) tuples keep the last
+// value. Numeric cells populate Value, others ValueText. A year column's
+// EndDate (nil for a point-in-time mapping) carries through to the observation
+// unchanged.
 func buildObservations(datasetID uuid.UUID, katoColumn string, years []domain.YearColumn, rows []map[string]string) []domain.Observation {
 	obs := make([]domain.Observation, 0, len(rows)*len(years))
 	index := make(map[string]int)
@@ -761,7 +774,7 @@ func buildObservations(datasetID uuid.UUID, katoColumn string, years []domain.Ye
 			continue
 		}
 		for _, yc := range years {
-			o := domain.Observation{DatasetID: datasetID, KatoCode: kato, Date: yc.Date, DateTo: yc.DateTo}
+			o := domain.Observation{DatasetID: datasetID, KatoCode: kato, StartDate: yc.StartDate, EndDate: yc.EndDate}
 			if raw := strings.TrimSpace(row[yc.Column]); raw != "" {
 				if f, err := strconv.ParseFloat(raw, 64); err == nil {
 					o.Value = &f
@@ -769,7 +782,7 @@ func buildObservations(datasetID uuid.UUID, katoColumn string, years []domain.Ye
 					o.ValueText = &raw
 				}
 			}
-			key := kato + "\x00" + yc.Date + "\x00" + dateToKey(yc.DateTo)
+			key := kato + "\x00" + yc.StartDate + "\x00" + endDateKey(yc.EndDate)
 			if i, ok := index[key]; ok {
 				obs[i] = o
 			} else {
@@ -781,12 +794,12 @@ func buildObservations(datasetID uuid.UUID, katoColumn string, years []domain.Ye
 	return obs
 }

-// dateToKey normalizes an optional period end date for use in a dedup key.
-func dateToKey(dateTo *string) string {
-	if dateTo == nil {
+// endDateKey normalizes an optional period end date for use in a dedup key.
+func endDateKey(endDate *string) string {
+	if endDate == nil {
 		return ""
 	}
-	return *dateTo
+	return *endDate
 }

 // ObservationPage is a page of observations with pagination metadata.
@@ -888,7 +901,7 @@ func (s *DatasetService) generateGeoJSON(ctx context.Context, id uuid.UUID) erro
 // boundary taken from the districts table and its observation values nested
 // under a `data` object (keyed by date) alongside `kato` and `name`. KATO codes
 // with no matching district are skipped. When any observation carries a period
-// (DateTo set), its end date is additionally exposed under a `periods` object
+// (EndDate set), its end date is additionally exposed under a `periods` object
 // (keyed by the same start date used in `data`); `periods` is omitted entirely
 // when every observation is point-in-time, so a dataset mapped without periods
 // produces byte-identical output to before periods existed.
@@ -954,7 +967,7 @@ func (s *DatasetService) loadGeoJSONDataset(ctx context.Context, id uuid.UUID) (
 // the grouped observation values (keyed by date) nested under a `data` object
 // alongside `kato` and `name` in each Feature's properties. When periods holds
 // an entry for the KATO (at least one observation had a period), its date->
-// date_to map is additionally nested under a `periods` object; a KATO with no
+// end_date map is additionally nested under a `periods` object; a KATO with no
 // period observations gets no `periods` key at all. KATO codes with no
 // matching district are skipped. order drives the deterministic feature order.
 func (s *DatasetService) districtFeatures(ctx context.Context, grouped map[string]map[string]any, periods map[string]map[string]string, order []string) ([]domain.Feature, error) {
@@ -1018,8 +1031,8 @@ func vectorFeatureProperties(raw json.RawMessage) map[string]any {

 // groupObservationsByKato groups observations by KATO code into date->value
 // maps. The value is the numeric Value when present, otherwise ValueText,
-// otherwise nil (an empty cell). It also groups, per KATO, a date->date_to map
-// for observations that carry a period (DateTo set); a KATO with no period
+// otherwise nil (an empty cell). It also groups, per KATO, a start_date->end_date map
+// for observations that carry a period (EndDate set); a KATO with no period
 // observations gets no entry in periods at all, so callers can tell "no
 // periods" apart from "empty periods". order lists KATO codes in first-seen
 // order so the resulting feature order is deterministic.
@@ -1034,15 +1047,15 @@ func groupObservationsByKato(obs []domain.Observation) (grouped map[string]map[s
 			grouped[o.KatoCode] = values
 			order = append(order, o.KatoCode)
 		}
-		values[o.Date] = observationValue(o)
+		values[o.StartDate] = observationValue(o)

-		if o.DateTo != nil {
+		if o.EndDate != nil {
 			p, ok := periods[o.KatoCode]
 			if !ok {
 				p = make(map[string]string)
 				periods[o.KatoCode] = p
 			}
-			p[o.Date] = *o.DateTo
+			p[o.StartDate] = *o.EndDate
 		}
 	}
 	return grouped, periods, order
@@ -1074,6 +1087,10 @@ type DatasetUpdateInput struct {
 	Name        string
 	Description *string
 	Unit        *string
+	// DatasetGroupID, StartDate and EndDate are nullable: nil clears them.
+	DatasetGroupID *uuid.UUID
+	StartDate      *string
+	EndDate        *string
 }

 // Update replaces a dataset's editable metadata (category, name, description,
@@ -1082,7 +1099,18 @@ func (s *DatasetService) Update(ctx context.Context, id uuid.UUID, in DatasetUpd
 	if err := s.ensureCategoryExists(ctx, in.CategoryID); err != nil {
 		return domain.Dataset{}, err
 	}
-	return s.repo.UpdateFields(ctx, id, in.CategoryID, in.Name, in.Description, in.Unit)
+	if err := domain.ValidateDatePeriod(in.StartDate, in.EndDate); err != nil {
+		return domain.Dataset{}, err
+	}
+	return s.repo.UpdateFields(ctx, id, domain.DatasetFields{
+		CategoryID:     in.CategoryID,
+		Name:           in.Name,
+		Description:    in.Description,
+		Unit:           in.Unit,
+		DatasetGroupID: in.DatasetGroupID,
+		StartDate:      in.StartDate,
+		EndDate:        in.EndDate,
+	})
 }

 // DatasetPatchInput carries a sparse set of metadata field changes. A nil
@@ -1096,6 +1124,13 @@ type DatasetPatchInput struct {
 	DescriptionSet bool
 	Unit           *string
 	UnitSet        bool
+
+	DatasetGroupID    *uuid.UUID
+	DatasetGroupIDSet bool
+	StartDate         *string
+	StartDateSet      bool
+	EndDate           *string
+	EndDateSet        bool
 }

 // Patch applies a sparse update to a dataset's editable metadata (category,
@@ -1125,8 +1160,31 @@ func (s *DatasetService) Patch(ctx context.Context, id uuid.UUID, in DatasetPatc
 	if in.UnitSet {
 		unit = in.Unit
 	}
+	groupID := current.DatasetGroupID
+	if in.DatasetGroupIDSet {
+		groupID = in.DatasetGroupID
+	}
+	startDate := current.StartDate
+	if in.StartDateSet {
+		startDate = in.StartDate
+	}
+	endDate := current.EndDate
+	if in.EndDateSet {
+		endDate = in.EndDate
+	}
+	if err := domain.ValidateDatePeriod(startDate, endDate); err != nil {
+		return domain.Dataset{}, err
+	}

-	return s.repo.UpdateFields(ctx, id, categoryID, name, description, unit)
+	return s.repo.UpdateFields(ctx, id, domain.DatasetFields{
+		CategoryID:     categoryID,
+		Name:           name,
+		Description:    description,
+		Unit:           unit,
+		DatasetGroupID: groupID,
+		StartDate:      startDate,
+		EndDate:        endDate,
+	})
 }

 // ensureCategoryExists validates that categoryID references an existing
--- a/internal/service/dataset_test.go
+++ b/internal/service/dataset_test.go
@@ -56,19 +56,60 @@ func (r *stubDatasetRepo) GetByID(_ context.Context, id uuid.UUID) (domain.Datas
 	return d, nil
 }

-func (r *stubDatasetRepo) UpdateFields(_ context.Context, id, categoryID uuid.UUID, name string, description, unit *string) (domain.Dataset, error) {
+func (r *stubDatasetRepo) UpdateFields(_ context.Context, id uuid.UUID, f domain.DatasetFields) (domain.Dataset, error) {
 	d, ok := r.store[id]
 	if !ok {
 		return domain.Dataset{}, domain.ErrNotFound
 	}
-	d.CategoryID = categoryID
-	d.Name = name
-	d.Description = description
-	d.Unit = unit
+	d.CategoryID = f.CategoryID
+	d.Name = f.Name
+	d.Description = f.Description
+	d.Unit = f.Unit
+	d.DatasetGroupID = f.DatasetGroupID
+	d.StartDate = f.StartDate
+	d.EndDate = f.EndDate
 	r.store[id] = d
 	return d, nil
 }

+func (r *stubDatasetRepo) LatestInGroup(_ context.Context, groupID uuid.UUID, status *string, fileType *domain.FileType) (domain.Dataset, error) {
+	var best *domain.Dataset
+	for _, d := range r.store {
+		d := d
+		if d.DatasetGroupID == nil || *d.DatasetGroupID != groupID {
+			continue
+		}
+		if status != nil && d.Status != *status {
+			continue
+		}
+		if fileType != nil && d.FileType != *fileType {
+			continue
+		}
+		if best == nil || laterPeriod(d, *best) {
+			best = &d
+		}
+	}
+	if best == nil {
+		return domain.Dataset{}, domain.ErrNotFound
+	}
+	return *best, nil
+}
+
+// laterPeriod reports whether a is fresher than b: greater end_date, NULL
+// end_date last, then newer created_at.
+func laterPeriod(a, b domain.Dataset) bool {
+	switch {
+	case a.EndDate != nil && b.EndDate == nil:
+		return true
+	case a.EndDate == nil && b.EndDate != nil:
+		return false
+	case a.EndDate != nil && b.EndDate != nil && *a.EndDate != *b.EndDate:
+		return *a.EndDate > *b.EndDate
+	default:
+		return a.CreatedAt.After(b.CreatedAt)
+	}
+}
+
 func (r *stubDatasetRepo) ListSummaries(_ context.Context, filter domain.DatasetFilter, limit, offset int) ([]domain.DatasetSummary, error) {
 	r.lastFilter = filter
 	r.lastLimit = limit
@@ -756,7 +797,7 @@ func TestDatasetService_SaveMapping(t *testing.T) {
 		Status:           domain.DatasetStatusAwaitingMapping,
 		AttributeColumns: []domain.AttributeColumn{{Name: "като"}, {Name: "F_2023"}},
 	}
-	validYears := []domain.YearColumn{{Column: "F_2023", Date: "2023-01-01"}}
+	validYears := []domain.YearColumn{{Column: "F_2023", StartDate: "2023-01-01"}}

 	newSvc := func() (*stubDatasetRepo, *DatasetService) {
 		repo := newStubDatasetRepo()
@@ -774,31 +815,31 @@ func TestDatasetService_SaveMapping(t *testing.T) {

 	t.Run("unknown year column", func(t *testing.T) {
 		_, svc := newSvc()
-		_, err := svc.SaveMapping(ctx, id, MappingInput{KatoColumn: "като", YearColumns: []domain.YearColumn{{Column: "X", Date: "2023-01-01"}}})
+		_, err := svc.SaveMapping(ctx, id, MappingInput{KatoColumn: "като", YearColumns: []domain.YearColumn{{Column: "X", StartDate: "2023-01-01"}}})
 		if !errors.Is(err, domain.ErrValidation) {
 			t.Fatalf("want ErrValidation, got %v", err)
 		}
 	})

-	t.Run("bad date", func(t *testing.T) {
+	t.Run("bad start_date", func(t *testing.T) {
 		_, svc := newSvc()
-		_, err := svc.SaveMapping(ctx, id, MappingInput{KatoColumn: "като", YearColumns: []domain.YearColumn{{Column: "F_2023", Date: "2023"}}})
+		_, err := svc.SaveMapping(ctx, id, MappingInput{KatoColumn: "като", YearColumns: []domain.YearColumn{{Column: "F_2023", StartDate: "2023"}}})
 		if !errors.Is(err, domain.ErrValidation) {
 			t.Fatalf("want ErrValidation, got %v", err)
 		}
 	})

-	t.Run("bad date_to", func(t *testing.T) {
+	t.Run("bad end_date", func(t *testing.T) {
 		_, svc := newSvc()
-		_, err := svc.SaveMapping(ctx, id, MappingInput{KatoColumn: "като", YearColumns: []domain.YearColumn{{Column: "F_2023", Date: "2023-01-01", DateTo: strPtr("2023")}}})
+		_, err := svc.SaveMapping(ctx, id, MappingInput{KatoColumn: "като", YearColumns: []domain.YearColumn{{Column: "F_2023", StartDate: "2023-01-01", EndDate: strPtr("2023")}}})
 		if !errors.Is(err, domain.ErrValidation) {
 			t.Fatalf("want ErrValidation, got %v", err)
 		}
 	})

-	t.Run("date_to before date", func(t *testing.T) {
+	t.Run("end_date before start_date", func(t *testing.T) {
 		_, svc := newSvc()
-		_, err := svc.SaveMapping(ctx, id, MappingInput{KatoColumn: "като", YearColumns: []domain.YearColumn{{Column: "F_2023", Date: "2023-12-31", DateTo: strPtr("2023-01-01")}}})
+		_, err := svc.SaveMapping(ctx, id, MappingInput{KatoColumn: "като", YearColumns: []domain.YearColumn{{Column: "F_2023", StartDate: "2023-12-31", EndDate: strPtr("2023-01-01")}}})
 		if !errors.Is(err, domain.ErrValidation) {
 			t.Fatalf("want ErrValidation, got %v", err)
 		}
@@ -808,12 +849,12 @@ func TestDatasetService_SaveMapping(t *testing.T) {
 		repo, svc := newSvc()
 		got, err := svc.SaveMapping(ctx, id, MappingInput{
 			KatoColumn:  "като",
-			YearColumns: []domain.YearColumn{{Column: "F_2023", Date: "2020-01-01", DateTo: strPtr("2023-12-31")}},
+			YearColumns: []domain.YearColumn{{Column: "F_2023", StartDate: "2020-01-01", EndDate: strPtr("2023-12-31")}},
 		})
 		if err != nil {
 			t.Fatalf("unexpected error: %v", err)
 		}
-		if len(got.YearColumns) != 1 || got.YearColumns[0].DateTo == nil || *got.YearColumns[0].DateTo != "2023-12-31" {
+		if len(got.YearColumns) != 1 || got.YearColumns[0].EndDate == nil || *got.YearColumns[0].EndDate != "2023-12-31" {
 			t.Fatalf("period not saved: %+v", got.YearColumns)
 		}
 		if len(repo.store[id].YearColumns) != 1 {
@@ -866,8 +907,8 @@ func TestDatasetService_SaveMapping(t *testing.T) {
 func TestBuildObservations(t *testing.T) {
 	id := uuid.New()
 	years := []domain.YearColumn{
-		{Column: "F_2023", Date: "2023-01-01"},
-		{Column: "D_2025", Date: "2025-01-01"},
+		{Column: "F_2023", StartDate: "2023-01-01"},
+		{Column: "D_2025", StartDate: "2025-01-01"},
 	}
 	rows := []map[string]string{
 		{"като": "751010000", "F_2023": "100", "D_2025": "n/a"},
@@ -882,7 +923,7 @@ func TestBuildObservations(t *testing.T) {

 	byKey := map[string]domain.Observation{}
 	for _, o := range obs {
-		byKey[o.KatoCode+"|"+o.Date] = o
+		byKey[o.KatoCode+"|"+o.StartDate] = o
 	}
 	if o := byKey["751010000|2023-01-01"]; o.Value == nil || *o.Value != 100 {
 		t.Errorf("numeric cell not stored as value: %+v", o)
@@ -891,8 +932,8 @@ func TestBuildObservations(t *testing.T) {
 		t.Errorf("non-numeric cell not stored as value_text: %+v", o)
 	}
 	for _, o := range obs {
-		if o.DateTo != nil {
-			t.Fatalf("point-in-time mapping must not set DateTo, got %+v", o)
+		if o.EndDate != nil {
+			t.Fatalf("point-in-time mapping must not set EndDate, got %+v", o)
 		}
 	}
 }
@@ -900,8 +941,8 @@ func TestBuildObservations(t *testing.T) {
 func TestBuildObservations_Period(t *testing.T) {
 	id := uuid.New()
 	years := []domain.YearColumn{
-		{Column: "F_2020_2023", Date: "2020-01-01", DateTo: strPtr("2023-12-31")}, // period
-		{Column: "F_2024", Date: "2024-01-01"},                                    // point-in-time, unaffected
+		{Column: "F_2020_2023", StartDate: "2020-01-01", EndDate: strPtr("2023-12-31")}, // period
+		{Column: "F_2024", StartDate: "2024-01-01"},                                     // point-in-time, unaffected
 	}
 	rows := []map[string]string{
 		{"като": "751010000", "F_2020_2023": "100", "F_2024": "200"},
@@ -912,22 +953,22 @@ func TestBuildObservations_Period(t *testing.T) {
 		t.Fatalf("want 2 observations, got %d", len(obs))
 	}

-	byDate := map[string]domain.Observation{}
+	byStartDate := map[string]domain.Observation{}
 	for _, o := range obs {
-		byDate[o.Date] = o
+		byStartDate[o.StartDate] = o
 	}

-	period := byDate["2020-01-01"]
-	if period.DateTo == nil || *period.DateTo != "2023-12-31" {
-		t.Fatalf("period observation missing date_to: %+v", period)
+	period := byStartDate["2020-01-01"]
+	if period.EndDate == nil || *period.EndDate != "2023-12-31" {
+		t.Fatalf("period observation missing end_date: %+v", period)
 	}
 	if period.Value == nil || *period.Value != 100 {
 		t.Fatalf("period observation value wrong: %+v", period)
 	}

-	point := byDate["2024-01-01"]
-	if point.DateTo != nil {
-		t.Fatalf("point-in-time observation must have nil date_to (backward compat), got %+v", point)
+	point := byStartDate["2024-01-01"]
+	if point.EndDate != nil {
+		t.Fatalf("point-in-time observation must have nil end_date (backward compat), got %+v", point)
 	}
 }

@@ -940,7 +981,7 @@ func TestDatasetService_Extract(t *testing.T) {
 		Filename: "r.geojson", StorageKey: "k",
 		Status:      domain.DatasetStatusExtracting,
 		KatoColumn:  &kato,
-		YearColumns: []domain.YearColumn{{Column: "F_2023", Date: "2023-01-01"}},
+		YearColumns: []domain.YearColumn{{Column: "F_2023", StartDate: "2023-01-01"}},
 	}
 	repo.districts["751010000"] = domain.District{
 		Kato: "751010000", Name: "Almaty",
@@ -1002,9 +1043,9 @@ func TestDatasetService_GeoJSON_DistrictJoin(t *testing.T) {

 	v2020, v2021 := 100.0, 150.0
 	repo.observations[id] = []domain.Observation{
-		{KatoCode: "710000000", Date: "2020-01-01", Value: &v2020},
-		{KatoCode: "710000000", Date: "2021-01-01", Value: &v2021},
-		{KatoCode: "999999999", Date: "2020-01-01", Value: &v2020}, // no district -> skipped
+		{KatoCode: "710000000", StartDate: "2020-01-01", Value: &v2020},
+		{KatoCode: "710000000", StartDate: "2021-01-01", Value: &v2021},
+		{KatoCode: "999999999", StartDate: "2020-01-01", Value: &v2020}, // no district -> skipped
 	}
 	repo.districts["710000000"] = domain.District{
 		Kato: "710000000", Name: "Astana",
@@ -1050,8 +1091,8 @@ func TestDatasetService_GeoJSON_Period(t *testing.T) {

 	period, point := 100.0, 200.0
 	repo.observations[id] = []domain.Observation{
-		{KatoCode: "710000000", Date: "2020-01-01", DateTo: strPtr("2023-12-31"), Value: &period},
-		{KatoCode: "710000000", Date: "2024-01-01", Value: &point}, // point-in-time, no period
+		{KatoCode: "710000000", StartDate: "2020-01-01", EndDate: strPtr("2023-12-31"), Value: &period},
+		{KatoCode: "710000000", StartDate: "2024-01-01", Value: &point}, // point-in-time, no period
 	}
 	repo.districts["710000000"] = domain.District{
 		Kato: "710000000", Name: "Astana",
@@ -1092,7 +1133,7 @@ func TestDatasetService_GeoJSON_NoPeriods_OmitsPeriodsKey(t *testing.T) {
 	repo := newStubDatasetRepo()
 	repo.store[id] = domain.Dataset{ID: id, FileType: domain.FileTypeVectorWithKato, Status: domain.DatasetStatusReady}
 	v := 100.0
-	repo.observations[id] = []domain.Observation{{KatoCode: "710000000", Date: "2020-01-01", Value: &v}}
+	repo.observations[id] = []domain.Observation{{KatoCode: "710000000", StartDate: "2020-01-01", Value: &v}}
 	repo.districts["710000000"] = domain.District{
 		Kato: "710000000", Name: "Astana",
 		Geometry: json.RawMessage(`{"type":"Polygon","coordinates":[[[71,51],[72,51],[72,52],[71,51]]]}`),
@@ -1116,8 +1157,8 @@ func TestDatasetService_GeoJSON_IgnoresDatasetGeometry(t *testing.T) {
 	}
 	v := 100.0
 	repo.observations[id] = []domain.Observation{
-		{KatoCode: "710000000", Date: "2020-01-01", Value: &v},
-		{KatoCode: "999999999", Date: "2020-01-01", Value: &v}, // no district -> skipped
+		{KatoCode: "710000000", StartDate: "2020-01-01", Value: &v},
+		{KatoCode: "999999999", StartDate: "2020-01-01", Value: &v}, // no district -> skipped
 	}
 	district := json.RawMessage(`{"type":"Polygon","coordinates":[[[71,51],[72,51],[72,52],[71,51]]]}`)
 	repo.districts["710000000"] = domain.District{Kato: "710000000", Name: "Astana", Geometry: district}
@@ -1174,7 +1215,7 @@ func TestDatasetService_GeoJSON_LazilyGeneratesAndCaches(t *testing.T) {
 	repo := newStubDatasetRepo()
 	repo.store[id] = domain.Dataset{ID: id, FileType: domain.FileTypeVectorWithKato, Status: domain.DatasetStatusReady}
 	v := 100.0
-	repo.observations[id] = []domain.Observation{{KatoCode: "710000000", Date: "2020-01-01", Value: &v}}
+	repo.observations[id] = []domain.Observation{{KatoCode: "710000000", StartDate: "2020-01-01", Value: &v}}
 	repo.districts["710000000"] = domain.District{
 		Kato: "710000000", Name: "Astana",
 		Geometry: json.RawMessage(`{"type":"Polygon","coordinates":[[[71,51],[72,51],[72,52],[71,51]]]}`),
@@ -1694,3 +1735,40 @@ func TestDatasetService_Patch_NotFound(t *testing.T) {
 		t.Fatalf("want ErrNotFound, got %v", err)
 	}
 }
+
+func TestDatasetService_Upload_StoresPeriodAndGroup(t *testing.T) {
+	repo := newStubDatasetRepo()
+	svc := newDatasetService(repo, &stubStore{}, true)
+
+	gid := uuid.New()
+	start, end := "2020-01-01", "2020-12-31"
+	in := validUpload()
+	in.DatasetGroupID = &gid
+	in.StartDate = &start
+	in.EndDate = &end
+
+	ds, err := svc.Upload(context.Background(), in)
+	if err != nil {
+		t.Fatal(err)
+	}
+	if ds.DatasetGroupID == nil || *ds.DatasetGroupID != gid || *ds.StartDate != start || *ds.EndDate != end {
+		t.Fatalf("period/group not stored: %+v", ds)
+	}
+}
+
+func TestDatasetService_Upload_RejectsReversedPeriod(t *testing.T) {
+	store := &stubStore{}
+	svc := newDatasetService(newStubDatasetRepo(), store, true)
+
+	start, end := "2021-01-01", "2020-12-31"
+	in := validUpload()
+	in.StartDate = &start
+	in.EndDate = &end
+
+	if _, err := svc.Upload(context.Background(), in); !errors.Is(err, domain.ErrValidation) {
+		t.Fatalf("want ErrValidation, got %v", err)
+	}
+	if len(store.put) != 0 {
+		t.Fatalf("nothing should be uploaded on validation failure, got %v", store.put)
+	}
+}
--- a/internal/transport/http/dataset_handler.go
+++ b/internal/transport/http/dataset_handler.go
@@ -80,9 +80,9 @@ func (h *DatasetHandler) status(w http.ResponseWriter, r *http.Request) {
 }

 type yearColumnInput struct {
-	Column string  `json:"column" validate:"required"`
-	Date   string  `json:"date" validate:"required,datetime=2006-01-02"`
-	DateTo *string `json:"date_to" validate:"omitempty,datetime=2006-01-02"`
+	Column    string  `json:"column" validate:"required"`
+	StartDate string  `json:"start_date" validate:"required,datetime=2006-01-02"`
+	EndDate   *string `json:"end_date" validate:"omitempty,datetime=2006-01-02"`
 }

 type mappingRequest struct {
@@ -109,7 +109,7 @@ func (h *DatasetHandler) mapping(w http.ResponseWriter, r *http.Request) {

 	in := service.MappingInput{KatoColumn: req.KatoColumn}
 	for _, yc := range req.YearColumns {
-		in.YearColumns = append(in.YearColumns, domain.YearColumn{Column: yc.Column, Date: yc.Date, DateTo: yc.DateTo})
+		in.YearColumns = append(in.YearColumns, domain.YearColumn{Column: yc.Column, StartDate: yc.StartDate, EndDate: yc.EndDate})
 	}

 	dataset, err := h.svc.SaveMapping(r.Context(), id, in)
@@ -150,6 +150,17 @@ func (h *DatasetHandler) upload(w http.ResponseWriter, r *http.Request) {
 		return
 	}

+	// The dataset group is optional; the period is validated by the service.
+	var groupID *uuid.UUID
+	if v := optionalFormValue(r, "dg_id"); v != nil {
+		gid, err := uuid.Parse(*v)
+		if err != nil {
+			httputil.WriteError(w, http.StatusUnprocessableEntity, "dg_id must be a valid UUID")
+			return
+		}
+		groupID = &gid
+	}
+
 	file, header, err := r.FormFile("file")
 	if err != nil {
 		httputil.WriteError(w, http.StatusBadRequest, "file is required")
@@ -163,17 +174,20 @@ func (h *DatasetHandler) upload(w http.ResponseWriter, r *http.Request) {
 	}

 	dataset, err := h.svc.Upload(r.Context(), service.UploadInput{
-		CategoryID:  categoryID,
-		Name:        r.FormValue("name"),
-		Description: optionalFormValue(r, "description"),
-		Unit:        optionalFormValue(r, "unit"),
-		Meta:        meta,
-		Automated:   automated,
-		Filename:    header.Filename,
-		FileType:    fileType,
-		ContentType: contentType,
-		Size:        header.Size,
-		Reader:      file,
+		CategoryID:     categoryID,
+		Name:           r.FormValue("name"),
+		Description:    optionalFormValue(r, "description"),
+		Unit:           optionalFormValue(r, "unit"),
+		Meta:           meta,
+		Automated:      automated,
+		Filename:       header.Filename,
+		FileType:       fileType,
+		ContentType:    contentType,
+		Size:           header.Size,
+		Reader:         file,
+		DatasetGroupID: groupID,
+		StartDate:      optionalFormValue(r, "start_date"),
+		EndDate:        optionalFormValue(r, "end_date"),
 	})
 	if err != nil {
 		respondDomainError(w, err)
@@ -226,7 +240,32 @@ func (h *DatasetHandler) list(w http.ResponseWriter, r *http.Request) {
 		categoryCode = &v
 	}

-	filter := domain.DatasetFilter{CategoryID: categoryID}
+	groupID, ok := parseOptionalUUIDQuery(w, r, "dg_id")
+	if !ok {
+		return
+	}
+
+	filter := domain.DatasetFilter{CategoryID: categoryID, DatasetGroupID: groupID}
+
+	for _, b := range []struct {
+		name string
+		dst  **string
+	}{
+		{"start_date_from", &filter.StartDateFrom},
+		{"start_date_to", &filter.StartDateTo},
+		{"end_date_from", &filter.EndDateFrom},
+		{"end_date_to", &filter.EndDateTo},
+	} {
+		v := strings.TrimSpace(r.URL.Query().Get(b.name))
+		if v == "" {
+			continue
+		}
+		if _, err := time.Parse("2006-01-02", v); err != nil {
+			httputil.WriteError(w, http.StatusBadRequest, b.name+" must be YYYY-MM-DD")
+			return
+		}
+		*b.dst = &v
+	}

 	if v := strings.TrimSpace(r.URL.Query().Get("file_type")); v != "" {
 		ft := domain.FileType(v)
@@ -320,10 +359,15 @@ type datasetPutRequest struct {
 	Name        string  `json:"name" validate:"required,max=255"`
 	Description *string `json:"description" validate:"required"`
 	Unit        *string `json:"unit" validate:"required"`
+	// DgID, StartDate and EndDate are optional; omitting them (or sending null
+	// or a blank string) clears the corresponding nullable column.
+	DgID      *string `json:"dg_id" validate:"omitempty,uuid"`
+	StartDate *string `json:"start_date" validate:"omitempty,datetime=2006-01-02"`
+	EndDate   *string `json:"end_date" validate:"omitempty,datetime=2006-01-02"`
 }

 // update replaces a dataset's editable metadata (category, name, description,
-// unit).
+// unit, dataset group, period).
 func (h *DatasetHandler) update(w http.ResponseWriter, r *http.Request) {
 	id, ok := parseUUIDParam(w, r, "id")
 	if !ok {
@@ -344,11 +388,24 @@ func (h *DatasetHandler) update(w http.ResponseWriter, r *http.Request) {
 		return
 	}

+	var groupID *uuid.UUID
+	if raw := optionalPtr(req.DgID); raw != nil {
+		gid, err := uuid.Parse(*raw)
+		if err != nil {
+			httputil.WriteError(w, http.StatusUnprocessableEntity, "dg_id must be a valid UUID")
+			return
+		}
+		groupID = &gid
+	}
+
 	dataset, err := h.svc.Update(r.Context(), id, service.DatasetUpdateInput{
-		CategoryID:  categoryID,
-		Name:        strings.TrimSpace(req.Name),
-		Description: normalizeOptional(*req.Description),
-		Unit:        normalizeOptional(*req.Unit),
+		CategoryID:     categoryID,
+		Name:           strings.TrimSpace(req.Name),
+		Description:    normalizeOptional(*req.Description),
+		Unit:           normalizeOptional(*req.Unit),
+		DatasetGroupID: groupID,
+		StartDate:      optionalPtr(req.StartDate),
+		EndDate:        optionalPtr(req.EndDate),
 	})
 	if err != nil {
 		respondDomainError(w, err)
@@ -360,11 +417,13 @@ func (h *DatasetHandler) update(w http.ResponseWriter, r *http.Request) {
 // datasetPatchFields lists the JSON keys accepted by PATCH /datasets/{id}.
 var datasetPatchFields = map[string]struct{}{
 	"category_id": {}, "name": {}, "description": {}, "unit": {},
+	"dg_id": {}, "start_date": {}, "end_date": {},
 }

 // patch applies a sparse update to a dataset's editable metadata (category,
-// name, description, unit): only fields present in the body are changed.
-// Sending an explicit null for description/unit clears that column.
+// name, description, unit, dg_id, start_date, end_date): only fields present in
+// the body are changed. Sending an explicit null for a nullable field clears
+// that column.
 func (h *DatasetHandler) patch(w http.ResponseWriter, r *http.Request) {
 	id, ok := parseUUIDParam(w, r, "id")
 	if !ok {
@@ -421,6 +480,35 @@ func (h *DatasetHandler) patch(w http.ResponseWriter, r *http.Request) {
 				return
 			}
 			in.Unit, in.UnitSet = unit, true
+		case "dg_id":
+			raw, err := decodeOptionalPatchString(val)
+			if err != nil {
+				httputil.WriteError(w, http.StatusUnprocessableEntity, "dg_id must be a valid UUID or null")
+				return
+			}
+			if raw != nil {
+				gid, err := uuid.Parse(*raw)
+				if err != nil {
+					httputil.WriteError(w, http.StatusUnprocessableEntity, "dg_id must be a valid UUID or null")
+					return
+				}
+				in.DatasetGroupID = &gid
+			}
+			in.DatasetGroupIDSet = true
+		case "start_date":
+			date, err := decodeOptionalPatchDate(val)
+			if err != nil {
+				httputil.WriteError(w, http.StatusUnprocessableEntity, "start_date must be YYYY-MM-DD or null")
+				return
+			}
+			in.StartDate, in.StartDateSet = date, true
+		case "end_date":
+			date, err := decodeOptionalPatchDate(val)
+			if err != nil {
+				httputil.WriteError(w, http.StatusUnprocessableEntity, "end_date must be YYYY-MM-DD or null")
+				return
+			}
+			in.EndDate, in.EndDateSet = date, true
 		}
 	}

@@ -432,6 +520,15 @@ func (h *DatasetHandler) patch(w http.ResponseWriter, r *http.Request) {
 	httputil.WriteJSON(w, http.StatusOK, dataset)
 }

+// optionalPtr normalizes an optional request field: a nil or blank value
+// yields nil, so an omitted or empty field clears a nullable column.
+func optionalPtr(v *string) *string {
+	if v == nil {
+		return nil
+	}
+	return normalizeOptional(*v)
+}
+
 // normalizeOptional trims s and returns nil when blank, so an empty string
 // clears a nullable column rather than being stored verbatim.
 func normalizeOptional(s string) *string {
@@ -455,6 +552,20 @@ func decodeOptionalPatchString(raw json.RawMessage) (*string, error) {
 	return normalizeOptional(s), nil
 }

+// decodeOptionalPatchDate unmarshals a PATCH date field that may be a
+// YYYY-MM-DD JSON string or null, returning nil (clear the column) for null or
+// a blank string.
+func decodeOptionalPatchDate(raw json.RawMessage) (*string, error) {
+	v, err := decodeOptionalPatchString(raw)
+	if err != nil || v == nil {
+		return nil, err
+	}
+	if _, err := time.Parse("2006-01-02", *v); err != nil {
+		return nil, err
+	}
+	return v, nil
+}
+
 // geojson returns the dataset's pre-assembled GeoJSON FeatureCollection (RFC
 // 7946), generated and stored at processing time. A vector_with_kato dataset's
 // collection joins the districts table on KATO code (one feature per KATO with
--- a/migrations/00010_dataset_groups_and_dates.sql
+++ b/migrations/00010_dataset_groups_and_dates.sql
@@ -0,0 +1,67 @@
+-- +goose Up
+-- Dataset groups: "одноимённые" datasets versioned by period. A group may carry a
+-- shared SLD used as a fallback for its datasets.
+CREATE TABLE dataset_groups (
+    id               UUID PRIMARY KEY      DEFAULT gen_random_uuid(),
+    code             VARCHAR(255) NOT NULL UNIQUE,
+    name             VARCHAR(255) NOT NULL,
+    description      TEXT         NOT NULL DEFAULT '',
+    sld_storage_key  TEXT,
+    sld_filename     VARCHAR(255),
+    sld_content_type VARCHAR(255),
+    created_at       TIMESTAMPTZ  NOT NULL DEFAULT now(),
+    updated_at       TIMESTAMPTZ  NOT NULL DEFAULT now()
+);
+
+ALTER TABLE datasets ADD COLUMN dg_id      UUID REFERENCES dataset_groups (id) ON DELETE RESTRICT;
+ALTER TABLE datasets ADD COLUMN start_date DATE;
+ALTER TABLE datasets ADD COLUMN end_date   DATE;
+ALTER TABLE datasets ADD CONSTRAINT chk_datasets_period
+    CHECK (start_date IS NULL OR end_date IS NULL OR end_date >= start_date);
+CREATE INDEX idx_datasets_dg_id_end_date ON datasets (dg_id, end_date DESC NULLS LAST, created_at DESC);
+CREATE INDEX idx_datasets_start_date ON datasets (start_date);
+CREATE INDEX idx_datasets_end_date   ON datasets (end_date);
+
+-- Period columns are renamed everywhere: observations, and the year_columns
+-- mapping JSON ({column, date, date_to} -> {column, start_date, end_date}).
+ALTER TABLE dataset_observations RENAME COLUMN date    TO start_date;
+ALTER TABLE dataset_observations RENAME COLUMN date_to TO end_date;
+DROP INDEX uq_dataset_observations;
+DROP INDEX idx_dataset_observations_kato_date;
+CREATE UNIQUE INDEX uq_dataset_observations ON dataset_observations (dataset_id, kato_code, start_date, COALESCE(end_date, start_date));
+CREATE INDEX idx_dataset_observations_kato_date ON dataset_observations (kato_code, start_date);
+
+UPDATE datasets
+SET year_columns = (
+    SELECT jsonb_agg(
+        (elem - 'date' - 'date_to')
+        || jsonb_build_object('start_date', elem->'date')
+        || CASE WHEN elem ? 'date_to' AND elem->'date_to' IS NOT NULL
+                THEN jsonb_build_object('end_date', elem->'date_to') ELSE '{}'::jsonb END)
+    FROM jsonb_array_elements(year_columns) AS elem)
+WHERE year_columns IS NOT NULL AND jsonb_typeof(year_columns) = 'array' AND jsonb_array_length(year_columns) > 0;
+
+-- +goose Down
+UPDATE datasets
+SET year_columns = (
+    SELECT jsonb_agg(
+        (elem - 'start_date' - 'end_date')
+        || jsonb_build_object('date', elem->'start_date')
+        || CASE WHEN elem ? 'end_date' AND elem->'end_date' IS NOT NULL
+                THEN jsonb_build_object('date_to', elem->'end_date') ELSE '{}'::jsonb END)
+    FROM jsonb_array_elements(year_columns) AS elem)
+WHERE year_columns IS NOT NULL AND jsonb_typeof(year_columns) = 'array' AND jsonb_array_length(year_columns) > 0;
+DROP INDEX uq_dataset_observations;
+DROP INDEX idx_dataset_observations_kato_date;
+ALTER TABLE dataset_observations RENAME COLUMN start_date TO date;
+ALTER TABLE dataset_observations RENAME COLUMN end_date   TO date_to;
+CREATE UNIQUE INDEX uq_dataset_observations ON dataset_observations (dataset_id, kato_code, date, COALESCE(date_to, date));
+CREATE INDEX idx_dataset_observations_kato_date ON dataset_observations (kato_code, date);
+DROP INDEX idx_datasets_dg_id_end_date;
+DROP INDEX idx_datasets_start_date;
+DROP INDEX idx_datasets_end_date;
+ALTER TABLE datasets DROP CONSTRAINT chk_datasets_period;
+ALTER TABLE datasets DROP COLUMN dg_id;
+ALTER TABLE datasets DROP COLUMN start_date;
+ALTER TABLE datasets DROP COLUMN end_date;
+DROP TABLE dataset_groups;