Json in Postgres

Let’s say you have a json object stored in a table called process:

{
  "f33d3b3e-0b0b-4b0b-8b0b-0b0b0b0b0b0b": {
    "candidates": 1,
    "affected": 1,
    "skipped": 0,
    "failed": 0
  },
  "f22d3b3e-0b0b-4b0b-8b0b-0b0b0b0b0b0b": {
    "candidates": 22,
    "affected": 22,
    "skipped": 0,
    "failed": 0
  }
}

Get each object as follows:

SELECT jsonb_each(stats) as entry from process;

returns a single column for the entry containing the key and value as a pair:

                                                   entry
----------------------------------------------------------------------------------------------------------------
 (f22d3b3e-0b0b-4b0b-8b0b-0b0b0b0b0b0b,"{""failed"": 0, ""skipped"": 0, ""affected"": 12, ""candidates"": 12}")
 (f33d3b3e-0b0b-4b0b-8b0b-0b0b0b0b0b0b,"{""failed"": 1, ""skipped"": 0, ""affected"": 10, ""candidates"": 11}")
(2 rows)

Or:

SELECT entry.* from process, jsonb_each(stats) as entry ;

returns separate columns for the key and the value:

                 key                  |                             value
--------------------------------------+---------------------------------------------------------------
 f22d3b3e-0b0b-4b0b-8b0b-0b0b0b0b0b0b | {"failed": 0, "skipped": 0, "affected": 12, "candidates": 12}
 f33d3b3e-0b0b-4b0b-8b0b-0b0b0b0b0b0b | {"failed": 1, "skipped": 0, "affected": 10, "candidates": 11}
(2 rows)

Let’s say we want to sum thet candidates in the value of the entry

SELECT sum((value->>'candidates')::int) from process, jsonb_each(stats) as x;

Reference

What operators to use e.g. -> vs ->>, see: https://www.postgresql.org/docs/9.5/functions-json.html

---

• Tags: postgres json sql