postgresql-interfaces/psqlodbc GitHub issues and pull requests (mirror)  

help / color / mirror / Atom feed

From: raddakal (@raddakal) <[email protected]>
To: postgresql-interfaces/psqlodbc <[email protected]>
Subject: [postgresql-interfaces/psqlodbc] issue #98: SELECT of a PostgreSQL CHAR/VARCHAR/TEXT Column Using multiple calls SQLGetData results in data truncation (last byte)
Date: Tue, 18 Mar 2025 05:44:32 +0000
Message-ID: <[email protected]> (raw)

When We execute the SELECT query on a table containing a TEXT column with one record of size 1048000 bytes, our ODBC application is able to retrieve one byte less than the actual data.

DDL and DML used in the test
------------------------------

postgres=# create table public.rantexttab(textcol text);
CREATE TABLE

postgres=# insert into public.rantexttab values(repeat('A', 1048000) || 'BB');
INSERT 0 1
postgres=# select length(textcol) from public.rantexttab;
 length  
---------
 1048002
(1 row)


Please note that We are using SQL_C_BINARY as the c type to receive the text column data.

We had allocated a buffer of 1048000 in our ODBC program and the size of teh actual data was 1048002, so 2 subsequent SQLGetData calls were made. In the first call, driver returned correctly 1048000 bytes and the in the second call  with the same buffer, driver returns two bytes the last byte contained a null (instead of teh actual last byte in the data).
 

When we did some more analysis, it seems multiple calls to SQLGetData is causing the issue (i.e., there seems to be some gap in the driver code when multiple calls to SQLGetData are made to retrieve the data of the same column in the result set).


Tried to shorten the test case, and instead of using the size of 1048002 bytes, we could reproduce the same issue using 12 bytes of data in TEXT column. When a buffer of size 10 bytes is used and 2 calls to SQLGetData being made, in the second call we could see the last byte was null instead of actual last byte in the column data.

In the shortened test case, the data used:

postgres=# select length(textcol) from public.rantexttab;
 length 
--------
     12
(1 row)

postgres=# SELECT textcol from public.rantexttab
postgres-# ;
   textcol    
--------------
 AAAAAAAAAABB
(1 row)




Standalone ODBC program output:
----------------------------------

bash-4.4$ ./a.out 
Connecting to database.
Connected to database.
Fetching results...
First Call to GetData: 
ind : 12
Data: AAAAAAAAAA

Second Call to GetData: 
ind : 2
Data: B
Free handles.


We can see that in the second call instead of getting BB, we get only B.



Thanks.

[pglob.cpp.txt](https://github.com/user-attachments/files/19307243/pglob.cpp.txt)

view thread (10+ messages)  latest in thread

reply

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Reply to all the recipients using the --to and --cc options:
  reply via email

  To: github://postgresql-interfaces/psqlodbc
  Cc: [email protected], [email protected]
  Subject: Re: [postgresql-interfaces/psqlodbc] issue #98: SELECT of a PostgreSQL CHAR/VARCHAR/TEXT Column Using multiple calls SQLGetData results in data truncation (last byte)
  In-Reply-To: <<[email protected]>>

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

This inbox is served by agora; see mirroring instructions
for how to clone and mirror all data and code used for this inbox