It looks like DOLT_DIFF requires a single table name, so I'm not sure it makes sense to factor in dolt_ignore here.

DOLT_DIFF_STAT can diff all tables (and accepts "WORKING" as an input), but only shows stats for tables that exist in both revisions. Because dolt_ignore is only supposed to filter tables that were added in the working set, it shouldn't affect the behavior of DOLT_DIFF_STAT either.

So the only function we need to consider is DOLT_DIFF_SUMMARY. A repro would look like this:

Setup:

INSERT INTO dolt_ignore VALUE ('t', true);
CREATE TABLE t (pk int);
SELECT * FROM DOLT_DIFF_SUMMARY("HEAD", "WORKING");

Current output:

+-----------------+---------------+-----------+-------------+---------------+
| from_table_name | to_table_name | diff_type | data_change | schema_change |
+-----------------+---------------+-----------+-------------+---------------+
|                 | t             | added     | false       | true          |
+-----------------+---------------+-----------+-------------+---------------+

Expected output:

Empty set

Uh oh!

dolt_diff_summary system table should hide tables ignored by dolt_ignore #5861

Description

Metadata

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

Issue actions

`dolt_diff_summary` system table should hide tables ignored by `dolt_ignore` #5861