cannot version unversioned table (inconsistent view on the matter)

[Sandro Santilli]

See this session:

lol_3_17=# select table_version.ver_version();
1.7.0dev 1.6.0-16-g7fabb6f
lol_3_17=# select table_version.ver_is_table_versioned('bde_ext', 'adjustment_run');
f
lol_3_17=# select table_version.ver_enable_versioning('bde_ext', 'adjustment_run');
ERROR:  Table bde_ext.adjustment_run is already versioned

Basically, ver_is_table_versioned says a table is NOT versioned while ver_enable_versioning says the table is ALREADY versioned... annoying enough

[Sandro Santilli]

There's no trigger on the table:

lol_3_17=# \d bde_ext.adjustment_run
                Table "bde_ext.adjustment_run"
       Column       |            Type             | Modifiers 
--------------------+-----------------------------+-----------
 id                 | integer                     | not null
 adm_id             | integer                     | not null
 cos_id             | integer                     | not null
 status             | character varying(4)        | not null
 usr_id_exec        | character varying(20)       | not null
 adjust_datetime    | timestamp without time zone | 
 description        | character varying(100)      | 
 sum_sqrd_residuals | numeric(22,12)              | 
 redundancy         | numeric(22,12)              | 
 wrk_id             | integer                     | 
 audit_id           | integer                     | not null
Indexes:
    "pkey_adjustment_run" PRIMARY KEY, btree (id)

The table_version.versioned_table table is empty.
The only thing is that a table_version.bde_ext_adjustment_run_revision table exists.

What should we do in these cases ?
Shall we consider the table as versioned or non-versioned ?

[Ivan Mincik]

I think it should be considered as non-versioned. How has that happen ?

[Sandro Santilli]

I think it should be considered as non-versioned. How has that happen ?

I drop cascaded the schema containing those tables, and later re-created them via schema loader. The tables under `table_version` did not cascade (maybe this could be obtained, with some improvements to our DDL trigger).

[stale]

This issue has been automatically marked as stale as there has not been any activity for sometime. The issue will be closed in 14 days if no further activity.

[Ivan Mincik]

This task is still relevant.

[stale]

This issue has been automatically marked as stale as there has not been any activity for sometime. The issue will be closed in 14 days if no further activity.

[Ivan Mincik]

This task is still relevant.

[stale]

This issue has been automatically marked as stale as there has not been any activity for sometime. The issue will be closed in 14 days if no further activity.

[Ivan Mincik]

This task is still relevant.

[Bill Nelson]

This seems like quite an edge case. Would require someone to drop cascade and then rebuild with schema loader.