Datasets V1 (& earlier)

Datasets V3 is the current repository structure Kart uses - Datasets V1 is no longer supported except that Kart can upgrade an old repo from Datasets V1 to Datasets V3, using kart upgrade SOURCE DEST.

Nevertheless, documentation for Datasets V1 follows.

Repository Layout:

v0.2 with working copy

myproject/
  .git/                          # git stuff lives under here
    config                       # includes pointer to working-copy

  readme.txt                     # random file managed by git

  path/to/mydataset/             # dataset name is 'dataset' here
    legend.jpg                   # random file managed by git
    .sno-table/                  # features for mydataset live under here in git. Excluded in git WC
      sno.idxi                   # spatial index (optional)
      sno.idxd                   # spatial index (optional)
      meta/
        version
        gpkg_contents
        gpkg_metadata            # metadata content (optional)
        gpkg_metadata_reference  # metadata <> table pointers (optional)
        sqlite_table_info
        primary_key
        gpkg_geometry_columns
        gpkg_spatial_ref_sys
      1a/                        # sha1 of feature pk, first 2 hex chars
        2b/                      # sha1 of feature pk, second 2 hex chars
          bAsE64=                # feature (msgpack-encoded). Name is base64-encoded msgpack of PK value

  otherdataset/                  # dataset name is 'dataset' here
    .sno-table/                  # features for mydataset live under here in git. Excluded in git WC
    documentation.pdf            # random file managed by git

  myproject.gpkg                 # working copy for project. Ignored in git WC

v0.0

nz_primary_parcels_taranaki/
  meta/
    gpkg_contents
    gpkg_geometry_columns
    gpkg_metadata
    gpkg_metadata_reference
    gpkg_spatial_ref_sys
    sqlite_table_info
    version
  features/
    fffd/
      fffdf8fa-e14d-4ac9-a9ba-0bb9f263a69e/
        affected_surveys
        appellation
        calc_area
        fid
        geom
        id
        land_district
        parcel_intent
        statutory_actions
        survey_area
        titles
        topology_type

GPKG working copies (v0, v1 is similar)

How the __kxg_map (IDMAP) table works

state field:

  • 0 = unchanged

  • -1 = deleted

  • 1 = edited

TABLEfid

att

IDMAPuuid

fid

state

1

a

123

1

0

2

b

124

2

0

3

c

125

3

0

6

e

126

6

0

7

e

127

7

0

UPDATE t SET att=cc WHERE fid=3
INSERT INTO t (fid, att) VALUES (4, 'd')
DELETE FROM t WHERE fid=1

TABLEfid

att

IDMAPuuid

fid

state

.

.

123

1

-1

2

b

124

2

0

3

cc

125

3

1

4

d

NULL

4

0

6

e

126

6

0

7

e

127

7

0

UPDATE t SET fid=5 WHERE fid=2

TABLEfid

att

IDMAPuuid

fid

state

.

.

123

1

-1

5

b

124

5

1

3

cc

125

3

1

4

d

NULL

4

0

6

e

126

6

0

7

e

127

7

0

UPDATE t SET att=bb WHERE fid=5
UPDATE t SET att=ccc WHERE fid=3
UPDATE t SET att=dd WHERE fid=4

TABLEfid

att

IDMAPuuid

fid

state

.

.

123

1

-1

5

bb

124

5

1

3

ccc

125

3

1

4

dd

NULL

4

1

6

e

126

6

0

7

e

127

7

0

UPDATE t SET fid=2 WHERE fid=5
UPDATE t SET fid=1 WHERE fid=4
UPDATE t SET fid=9 WHERE fid=7

TABLEfid

att

IDMAPuuid

fid

state

.

.

123

1

-1

2

bb

124

2

1

3

ccc

125

3

1

1

dd

NULL

1

1

6

e

126

6

0

9

e

127

9

1

  • feature 123 was deleted

  • feature 124 was edited (att=b -> bb)

  • feature 125 was edited (att=c -> ccc)

  • feature fid=1 was added (fid= -> 1; att= -> dd)

  • feature 127 was edited (fid=7 -> 9)