Re: Proposal: pg_is_volatile function - Mailing list pgsql-hackers

From Pavel Stehule
Subject Re: Proposal: pg_is_volatile function
Date
Msg-id CAFj8pRD7SzbNF9WYnW-P8ebUbP-DrCynG=H4Y+KAimTnyBOU3Q@mail.gmail.com
Whole thread Raw
In response to Proposal: pg_is_volatile function  (Andrew Farries <andrew.farries@xata.io>)
Responses Re: Proposal: pg_is_volatile function
List pgsql-hackers
Hi

čt 20. 2. 2025 v 13:48 odesílatel Andrew Farries <andrew.farries@xata.io> napsal:
I'd like to propose a new function `pg_is_volatile` that would test and return
the volatility of its argument expression. Example uses of the function would
be:

pg_is_volatile(1) -> false
pg_is_volatile(random()) -> true

The motivation for the proposal is to allow testing of column default
expressions for new columns added with `ALTER TABLE ... ADD COLUMN` before
adding the column. This is to determine whether the column default will be able
to take advantage of the fast-path optimization for non-volatile column
defaults, or whether a full table rewrite will be required.

For a schema migration tool, it's desirable for the tool to assess the
volatility of a column default for a new column before adding it. The tool can
then decide on the most appropriate way to add the column, either doing so
directly for a non-volatile default, or issuing a warning or using some other
method in the case of a volatile default.

The documentation for this function would be as follows:

```
  <row>
   <entry role="func_table_entry"><para role="func_signature">
    <indexterm>
     <primary>pg_is_volatile</primary>
    </indexterm>
    <function>pg_is_volatile</function> ( <type>"any"</type> )
    <returnvalue>boolean</returnvalue>
   </para>
   <para>
    Tests whether the argument expression contains volatile functions (see
    <xref linkend="xfunc-volatility"/>). This can be useful to determine
    whether the expression can be used as a column default without causing
    a table rewrite.
   </para></entry>
  </row>
```

I believe the implementation of this function would be straightforward with a
new function in `src/backend/utils/adt/misc.c` delegating to the existing
`contain_volatile_functions_after_planning` function in
`src/backend/optimizer/util/clauses.c`.


If this feature can be implemented, then it needs to be implemented like a pseudo function, it cannot not be a classic function.

But for your use case you should probably check if the function is stable too? So maybe you should check if the expression is immutable.

Maybe this function can be designed like pg_get_expr_volatility(expr) and returns v, s or i

Probably this functionality should not be used for any other cases, so I can imagine different and maybe more generic solutions. We can introduce the function
`pg_does_relation_rewrite(sqlstr)`. Then the external tool can have less knowledge about pg internals, and doesn't need to parse the command to separate the expression.

Regards

Pavel




 

pgsql-hackers by date:

Previous
From: Teodor Sigaev
Date:
Subject: Improvement of var_eq_non_const()
Next
From: "Devulapalli, Raghuveer"
Date:
Subject: RE: SIMD optimization for list_sort