MIME-Version: 1.0
From: Roberto Mello <roberto.mello@gmail.com>
Date: Wed, 25 Mar 2026 13:50:45 -0600
Message-ID: 
 <CAKz==b+k7OwOib__LgfDadUzW9jFrbQ9FqdERpsufi8d7F=vkQ@mail.gmail.com>
Subject: pg_publication_tables: return NULL attnames when no column list is
 specified
To: PostgreSQL Developers <pgsql-hackers@lists.postgresql.org>
Content-Type: multipart/mixed; boundary="000000000000062be4064dde97ae"
Archived-At: 
 <https://www.postgresql.org/message-id/CAKz%3D%3Db%2Bk7OwOib__LgfDadUzW9jFrbQ9FqdERpsufi8d7F%3DvkQ%40mail.gmail.com>
Precedence: bulk

--000000000000062be4064dde97ae
Content-Type: multipart/alternative; boundary="000000000000062be3064dde97ac"

--000000000000062be3064dde97ac
Content-Type: text/plain; charset="UTF-8"

Hi,

When a table is added to a publication without a column list:

ALTER PUBLICATION my_pub ADD TABLE my_table;

Any new column added to my_table will automatically be replicated. The
underlying catalog (pg_publication_rel.prattrs) stores NULL for this case.

When a table is added with an explicit column list:

ALTER PUBLICATION my_pub ADD TABLE my_table (id, name, status);

New columns will NOT be replicated until the publication is explicitly
altered.
The catalog stores an int2vector of the specified attribute numbers.

The problem is that pg_get_publication_tables() in pg_publication.c (the
set-returning
function backing the pg_publication_tables view) erases this distinction.
When prattrs is NULL, it opens the table, iterates all eligible attributes,
and builds a synthetic int2vector of all current columns. The view then
shows
the same attnames output for both cases.

| Scenario                 | prattrs in catalog | attrs from SRF      |
attnames in view |
|--------------------------|--------------------|---------------------|------------------|
| column list (a, b)       | {1, 2  }           | {1,2}               | {a,
b}           |
| No column list           | NULL               | {1,2,3,...} (synth) | {a,
b, c, ...}   |
| FOR ALL TABLES/IN SCHEMA | no catalog row     | {1,2,3,...} (synth) | {a,
b, c, ...}   |

This is a problem for operations: schema migrations that add columns to
published
tables may or may not replicate depending on how the publication was
originally
defined, and the only way to check is querying pg_publication_rel directly.

Additionally, tablesync.c has a workaround heuristic that tries to reverse
this synthesis by comparing array_length(gpt.attrs, 1) against c.relnatts.
This heuristic is buggy: relnatts counts all user attributes including
dropped
columns, but the synthesized list excludes dropped columns and conditionally
excludes generated columns. So for any table with a dropped column, the
heuristic incorrectly treats "no column list" as if an explicit list were
specified.

The problematic code is the result of commits that incrementally
built on each other, with later ones introducing workarounds for side
effects
of earlier ones:

1. fd0b9dcebda (06/2022) Amit Kapila (author: Hou Zhijie)
Prohibit combining publications with different column lists

Added the subscriber-side check that prevents combining publications with
different column lists for the same table. The subscriber runs
SELECT DISTINCT gpt.attrs and expects at most one row. To handle the fact
that pg_get_publication_tables() returned NULL for no-column-list
publications (which would always be distinct from any explicit list), this
commit introduced the relnatts heuristic in tablesync.c:

CASE WHEN (array_length(gpt.attrs, 1) = c.relnatts)
     THEN NULL ELSE gpt.attrs END

The idea was: if the SRF returns a synthesized list whose length equals
relnatts, it must be the "all columns" case, so collapse it back to NULL.
But at this point the SRF didn't synthesize anything yet, so the heuristic
was anticipating the next commit.

2. b7ae0395369 (01/2023) Amit Kapila (author: Shi yu)
Ignore dropped and generated columns from the column list

This is the commit that introduced the synthesis loop. It "solved" a problem
where a table in two publications (one with a column list naming all
columns,
one without a column list) was erroring with "cannot use different column
lists"
because one returned NULL and the other returned an int2vector. The fix was
to synthesize a column list when none was specified, filtering out dropped
and generated columns, so that both publications would produce identical
int2vectors and DISTINCT would collapse them to one row.

But this synthesis made it impossible to distinguish "all columns" from
"explicit list of all columns" in the view. And the relnatts heuristic from
fd0b9dcebda - which was supposed to reverse the synthesis - was broken from
the start because relnatts includes dropped columns while the synthesized
list
excludes them.

The synthesis in b7ae0395369 tried to make "no column list" and "explicit
list
of all columns" look identical. But they have genuinely different
semantics:

- No column list (NULL): all current and future columns are replicated.
  ALTER TABLE ADD COLUMN automatically replicates the new column.
- Explicit full list: only the named columns are replicated. New columns
  are NOT replicated until the publication is explicitly altered.

By making them indistinguishable, the synthesis hid a real conflict from
users
who had a table in two publications with different column semantics on the
same subscription. I am proposing a fix that restores the distinction and
correctly
(IMO) surfaces this conflict.

The fix is to stop synthesizing the full column list in
pg_get_publication_tables().
When prattrs is NULL in the catalog, let attrs remain NULL in the SRF
output.
Remove the buggy CASE WHEN heuristic in tablesync.c since it is no longer
needed.

There is one scenario where there is a change for users: one pub no list +
one pub explicit all columns. Anyone with that specific configuration will
see a new error on the
next tablesync or subscription refresh after upgrading. The fix for those
users is to
either remove the explicit column list from the second publication (making
both "all
columns"), or keep the difference and use separate subscriptions.

AFAICT this is safe except for the change in behavior I describe above.
psql and pg_dump query pg_publication_rel directly and not the affected
view. New subscriber
talking to old publisher still works (old pub synthesizes a list, new sub
handles it).
Old subscriber talking to new publisher also works (the old CASE WHEN
evaluates
array_length(NULL, 1) which returns NULL, falling through to the ELSE branch
returning NULL).

Attached is a patch with the fix, including updates to documentation, 7
updated
regression tests where existing expected outputs where attnames changes from
{a} to NULL, and added new test cases that verify both cases are
distinguishable in
the same query.

Roberto Mello
Snowflake

--000000000000062be3064dde97ac
Content-Type: text/html; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

<div dir=3D"ltr">Hi,<br><br>When a table is added to a publication without =
a column list:<br><br>ALTER PUBLICATION my_pub ADD TABLE my_table;<br><br>A=
ny new column added to my_table will automatically be replicated. The<br>un=
derlying catalog (pg_publication_rel.prattrs) stores NULL for this case.<br=
><br>When a table is added with an explicit column list:<br><br>ALTER PUBLI=
CATION my_pub ADD TABLE my_table (id, name, status);<br><br>New columns wil=
l NOT be replicated until the publication is explicitly altered.<br>The cat=
alog stores an int2vector of the specified attribute numbers.<br><br>The pr=
oblem is that pg_get_publication_tables() in pg_publication.c (the set-retu=
rning<br>function backing the pg_publication_tables view) erases this disti=
nction.<br>When prattrs is NULL, it opens the table, iterates all eligible =
attributes,<br>and builds a synthetic int2vector of all current columns. Th=
e view then shows<br>the same attnames output for both cases.<br><br>| Scen=
ario =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 | prattrs in c=
atalog | attrs from SRF =C2=A0 =C2=A0 =C2=A0| attnames in view |<br>|------=
--------------------|--------------------|---------------------|-----------=
-------|<br>| column list (a, b) =C2=A0 =C2=A0 =C2=A0 | {1, 2 =C2=A0} =C2=
=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 | {1,2} =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =
=C2=A0 =C2=A0 | {a, b} =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 |<br>| No column =
list =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 | NULL =C2=A0 =C2=A0 =C2=A0 =C2=A0 =
=C2=A0 =C2=A0 =C2=A0 | {1,2,3,...} (synth) | {a, b, c, ...} =C2=A0 |<br>| F=
OR ALL TABLES/IN SCHEMA | no catalog row =C2=A0 =C2=A0 | {1,2,3,...} (synth=
) | {a, b, c, ...} =C2=A0 |<br><br>This is a problem for operations: schema=
 migrations that add columns to published<br>tables may or may not replicat=
e depending on how the publication was originally<br>defined, and the only =
way to check is querying pg_publication_rel directly.<br><br>Additionally, =
tablesync.c has a workaround heuristic that tries to reverse<br>this synthe=
sis by comparing array_length(gpt.attrs, 1) against c.relnatts.<br>This heu=
ristic is buggy: relnatts counts all user attributes including dropped<br>c=
olumns, but the synthesized list excludes dropped columns and conditionally=
<br>excludes generated columns. So for any table with a dropped column, the=
<br>heuristic incorrectly treats &quot;no column list&quot; as if an explic=
it list were<br>specified.<br><br>The problematic code is the result of com=
mits that incrementally<br>built on each other, with later ones introducing=
 workarounds for side effects<br>of earlier ones:<br><br>1. fd0b9dcebda (06=
/2022) Amit Kapila (author: Hou Zhijie)<br>Prohibit combining publications =
with different column lists<br><br>Added the subscriber-side check that pre=
vents combining publications with<br>different column lists for the same ta=
ble. The subscriber runs<br>SELECT DISTINCT gpt.attrs and expects at most o=
ne row. To handle the fact<br>that pg_get_publication_tables() returned NUL=
L for no-column-list<br>publications (which would always be distinct from a=
ny explicit list), this<br>commit introduced the relnatts heuristic in tabl=
esync.c:<br><br>CASE WHEN (array_length(gpt.attrs, 1) =3D c.relnatts)<br>=
=C2=A0 =C2=A0 =C2=A0THEN NULL ELSE gpt.attrs END<br><br>The idea was: if th=
e SRF returns a synthesized list whose length equals<br>relnatts, it must b=
e the &quot;all columns&quot; case, so collapse it back to NULL.<br>But at =
this point the SRF didn&#39;t synthesize anything yet, so the heuristic was=
 anticipating the next commit.<br><br>2. b7ae0395369 (01/2023) Amit Kapila =
(author: Shi yu)<br>Ignore dropped and generated columns from the column li=
st<br><br>This is the commit that introduced the synthesis loop. It &quot;s=
olved&quot; a problem<br>where a table in two publications (one with a colu=
mn list naming all columns,<br>one without a column list) was erroring with=
 &quot;cannot use different column lists&quot;<br>because one returned NULL=
 and the other returned an int2vector. The fix was<br>to synthesize a colum=
n list when none was specified, filtering out dropped<br>and generated colu=
mns, so that both publications would produce identical<br>int2vectors and D=
ISTINCT would collapse them to one row.<br><br>But this synthesis made it i=
mpossible to distinguish &quot;all columns&quot; from<br>&quot;explicit lis=
t of all columns&quot; in the view. And the relnatts heuristic from<br>fd0b=
9dcebda - which was supposed to reverse the synthesis - was broken from<br>=
the start because relnatts includes dropped columns while the synthesized l=
ist<br>excludes them.<br><br>The synthesis in b7ae0395369 tried to make &qu=
ot;no column list&quot; and &quot;explicit list<br>of all columns&quot; loo=
k identical. But they have genuinely different<br>semantics:<br><br>- No co=
lumn list (NULL): all current and future columns are replicated.<br>=C2=A0 =
ALTER TABLE ADD COLUMN automatically replicates the new column.<br>- Explic=
it full list: only the named columns are replicated. New columns<br>=C2=A0 =
are NOT replicated until the publication is explicitly altered.<br><br>By m=
aking them indistinguishable, the synthesis hid a real conflict from users<=
br>who had a table in two publications with different column semantics on t=
he<br>same subscription. I am proposing a fix that restores the distinction=
 and correctly<br>(IMO) surfaces this conflict.<br><br>The fix is to stop s=
ynthesizing the full column list in pg_get_publication_tables().<br>When pr=
attrs is NULL in the catalog, let attrs remain NULL in the SRF output.<br>R=
emove the buggy CASE WHEN heuristic in tablesync.c since it is no longer<br=
>needed.<br><br>There is one scenario where there is a change for users: on=
e pub no list + one pub explicit all columns.=C2=A0Anyone with that specifi=
c configuration will see a new error on the<div>next tablesync or subscript=
ion refresh after upgrading. The fix for those users is to=C2=A0</div><div>=
either remove the explicit column list from the second publication (making =
both &quot;all</div><div>columns&quot;), or keep the difference and use sep=
arate subscriptions.</div><div><div><br></div><div>AFAICT this is safe exce=
pt for the change in behavior I describe above. psql and pg_dump query pg_p=
ublication_rel directly and not the affected view. New subscriber</div><div=
>talking to old publisher still works (old pub synthesizes a list, new sub =
handles it).</div><div>Old subscriber talking to new publisher also works (=
the old CASE WHEN evaluates</div><div>array_length(NULL, 1) which returns N=
ULL, falling through to the ELSE branch</div><div>returning NULL).</div><di=
v><br></div><div>Attached is a patch with the fix, including updates to doc=
umentation, 7 updated</div><div>regression tests where existing expected ou=
tputs where attnames changes from</div><div>{a} to NULL, and added new test=
 cases that verify both cases are distinguishable in</div><div>the same que=
ry.</div><div><br></div><div>Roberto Mello</div><div>Snowflake</div></div><=
/div>

--000000000000062be3064dde97ac--

--000000000000062be4064dde97ae
Content-Type: application/octet-stream;
	name="v1-0001-Fix-pg_publication_tables-to-return-NULL-attnames.patch"
Content-Disposition: attachment;
	filename="v1-0001-Fix-pg_publication_tables-to-return-NULL-attnames.patch"
Content-Transfer-Encoding: base64
Content-ID: <f_mn6gfvkt0>
X-Attachment-Id: f_mn6gfvkt0
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--000000000000062be4064dde97ae--