Thread: SQL issue after migrating from version 13 to 15

SQL issue after migrating from version 13 to 15

From
"Campbell, Lance"
Date:

We migrated our PostgreSQL database from version 13 to 15. Absolutely no coding changes were made to our software.

table_column is a column in a table of type tsvector

 

The below segment of the where clause works fine if the value passed is a single value like “real”:

 

AND to_tsvector('simple', CAST (table_column as text)) @@ to_tsquery('simple', 'real')

 

However, this no longer works when there are two values “real,impact”. The only change was migrating from PostgreSQL 13 to 15:

 

AND to_tsvector('simple', CAST (table_column as text)) @@ to_tsquery('simple', 'real,impact')

 

No exception is being thrown.

 

Thanks for your help with this.

 

Lance Campbell

University of Illinois

Re: SQL issue after migrating from version 13 to 15

From
Tom Lane
Date:
"Campbell, Lance" <lance@illinois.edu> writes:
> The below segment of the where clause works fine if the value passed is a single value like "real":

> AND to_tsvector('simple', CAST (table_column as text)) @@ to_tsquery('simple', 'real')

> However, this no longer works when there are two values "real,impact". The only change was migrating from PostgreSQL
13to 15: 

> AND to_tsvector('simple', CAST (table_column as text)) @@ to_tsquery('simple', 'real,impact')

You really should define what you mean by "works" in a question
like this.

However, I think what you are unhappy about is that the interpretation
of that to_tsquery input has changed.  In v13:

regression=# select to_tsquery('simple', 'real,impact');
    to_tsquery
-------------------
 'real' & 'impact'
(1 row)

In v14 and later:

regression=# select to_tsquery('simple', 'real,impact');
     to_tsquery
---------------------
 'real' <-> 'impact'
(1 row)

The v14 release notes mention that there were incompatible changes in
this area, although they don't cite this specific case.  But anyway,
if the behavior you want is & then I'd suggest writing &, rather than
assuming that some other punctuation will behave the same.  Or you
could switch to plainto_tsquery(), which disregards the punctuation
altogether.

            regards, tom lane



RE: SQL issue after migrating from version 13 to 15

From
"Campbell, Lance"
Date:
Thanks so much for the assistance. That resolved my issue. I hope you have a great week.

Lance

-----Original Message-----
From: Tom Lane <tgl@sss.pgh.pa.us>
Sent: Monday, October 23, 2023 10:31 AM
To: Campbell, Lance <lance@illinois.edu>
Cc: pgsql-sql@lists.postgresql.org
Subject: Re: SQL issue after migrating from version 13 to 15

"Campbell, Lance" <lance@illinois.edu> writes:
> The below segment of the where clause works fine if the value passed is a single value like "real":

> AND to_tsvector('simple', CAST (table_column as text)) @@
> to_tsquery('simple', 'real')

> However, this no longer works when there are two values "real,impact". The only change was migrating from PostgreSQL
13to 15: 

> AND to_tsvector('simple', CAST (table_column as text)) @@
> to_tsquery('simple', 'real,impact')

You really should define what you mean by "works" in a question like this.

However, I think what you are unhappy about is that the interpretation of that to_tsquery input has changed.  In v13:

regression=# select to_tsquery('simple', 'real,impact');
    to_tsquery
-------------------
 'real' & 'impact'
(1 row)

In v14 and later:

regression=# select to_tsquery('simple', 'real,impact');
     to_tsquery
---------------------
 'real' <-> 'impact'
(1 row)

The v14 release notes mention that there were incompatible changes in this area, although they don't cite this specific
case. But anyway, if the behavior you want is & then I'd suggest writing &, rather than assuming that some other
punctuationwill behave the same.  Or you could switch to plainto_tsquery(), which disregards the punctuation
altogether.

            regards, tom lane