Calculated field
PARSE_WKT([wkt])
Sample data
| shipment | shipped | received | text | year | month | day | hour | minute | second | lat | lon | wkt |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 1 | 2024-12-29 | 2024-12-29T16:25:40 | 2024-12-29 | 2025 | 2 | 25 | 9 | 25 | 40 | 10 | 106 | POINT (106 10) |
| 2 | 2024-12-30 | 2024-12-30T16:25:40 | 2024-12-30 | 2025 | 2 | 26 | 10 | 25 | 40 | 10.1 | 106.1 | POINT (106.1 10.1) |
| 3 | 2025-01-01 | 2025-01-01T16:25:40 | 2025-01-01 | 2025 | 2 | 27 | 11 | 25 | 40 | 10.2 | 106.2 | POINT (106.2 10.2) |
| 4 | 2025-01-31 | 2025-01-31T16:25:40 | 2025-01-31 | 2025 | 2 | 28 | 12 | 25 | 40 | 10.3 | 106.3 | POINT (106.3 10.3) |
| 5 | 2025-02-28 | 2025-02-28T16:25:40 | 2025-02-28 | 2025 | 2 | 29 | 13 | 25 | 40 | 10.4 | 106.4 | POINT (106.4 10.4) |
| 6 | 2025-03-15 | 2025-03-15T16:25:40 | 2025-03-15 | 2025 | 2 | 30 | 14 | 25 | 40 | 10.5 | 106.5 | POINT (106.5 10.5) |
| 7 | null | null | null | 2025 | 2 | 31 | 15 | 25 | 40 | 10.6 | 106.6 | POINT (106.6 10.6) |
| 8 | 2024-02-29 | 2024-02-29T16:25:40 | 2024-02-29 | 2025 | 2 | 32 | 16 | 25 | 40 | 10.7 | 106.7 | POINT (106.7 10.7) |
The same calculation as SQL
SELECT json_object('type','Point','coordinates',json_array(regexp_extract(wkt,'POINT \(([^ ]+) ([^)]+)\)',1)::DOUBLE,regexp_extract(wkt,'POINT \(([^ ]+) ([^)]+)\)',2)::DOUBLE)) AS result FROM input ORDER BY shipmentRuns the example above.
SELECT ST_ASGEOJSON(TO_GEOGRAPHY(wkt)) AS result FROM input ORDER BY shipment
Written for this dialect; not run here.
SELECT ST_ASGEOJSON(ST_GEOGFROMTEXT(wkt)) AS result FROM input ORDER BY shipment
Written for this dialect; not run here.
Result
| result |
|---|
| {"type": "Point", "coordinates": [106, 10]} |
| {"type": "Point", "coordinates": [106.1, 10.1]} |
| {"type": "Point", "coordinates": [106.2, 10.2]} |
| {"type": "Point", "coordinates": [106.3, 10.3]} |
| {"type": "Point", "coordinates": [106.4, 10.4]} |
| {"type": "Point", "coordinates": [106.5, 10.5]} |
| {"type": "Point", "coordinates": [106.6, 10.6]} |
| {"type": "Point", "coordinates": [106.7, 10.7]} |
Same rows from the calculation and from the SQL in DuckDB.
Where a naive translation goes wrong
- WKT point coordinates are longitude followed by latitude.
Naive SQL
SELECT json_object('type','Point','coordinates',json_array(lat,lon)) AS result FROM input ORDER BY shipmentshipment shipped received text result 1 2024-12-29 2024-12-29T16:25:40 2024-12-29 {"type": "Point", "coordinates": [10, 106]} 2 2024-12-30 2024-12-30T16:25:40 2024-12-30 {"type": "Point", "coordinates": [10.1, 106.1]} 3 2025-01-01 2025-01-01T16:25:40 2025-01-01 {"type": "Point", "coordinates": [10.2, 106.2]} 4 2025-01-31 2025-01-31T16:25:40 2025-01-31 {"type": "Point", "coordinates": [10.3, 106.3]} 5 2025-02-28 2025-02-28T16:25:40 2025-02-28 {"type": "Point", "coordinates": [10.4, 106.4]} 6 2025-03-15 2025-03-15T16:25:40 2025-03-15 {"type": "Point", "coordinates": [10.5, 106.5]} 7 null null null {"type": "Point", "coordinates": [10.6, 106.6]} 8 2024-02-29 2024-02-29T16:25:40 2024-02-29 {"type": "Point", "coordinates": [10.7, 106.7]} Matches the calculation
SELECT json_object('type','Point','coordinates',json_array(regexp_extract(wkt,'POINT \(([^ ]+) ([^)]+)\)',1)::DOUBLE,regexp_extract(wkt,'POINT \(([^ ]+) ([^)]+)\)',2)::DOUBLE)) AS result FROM input ORDER BY shipmentresult {"type": "Point", "coordinates": [106, 10]} {"type": "Point", "coordinates": [106.1, 10.1]} {"type": "Point", "coordinates": [106.2, 10.2]} {"type": "Point", "coordinates": [106.3, 10.3]} {"type": "Point", "coordinates": [106.4, 10.4]} {"type": "Point", "coordinates": [106.5, 10.5]} {"type": "Point", "coordinates": [106.6, 10.6]} {"type": "Point", "coordinates": [106.7, 10.7]}
Signature
PARSE_WKT(string)