Re: pg_dump weirdness

That's correct.  It's backward compatible, not forward compatible.

On 11/5/21 3:59 PM, MichaelDBA@sqlexec.com wrote:
> You cannot use a newer version of pgdump to export and then insert it into an older version of PostgreSQL if I am
understandingyou correctly.
 
>
> Sent from my iPhone
>
>> On Nov 5, 2021, at 3:10 PM, John Scalia <jayknowsunix@gmail.com> wrote:
>>
>> I ran a pg_dump for a colleague, as he needs to move some data into a new database. I ran the command as:
>>
>> pg_dump -h original_server -t original_schema.table_name -U user -W dbname > table_name.sql
>>
>> after the dump was completed with no errors, I went into the target instance, and truncated the target table. I also
hadto do a sed operation to change the schema name in the dump file for the schema in the target. I verified that sed
properlymodified the file. Now, after all this, I used:
 
>>
>> psql -h target_server -U user newdbname < table_name.sql
>>
>> This loaded with a handful of warnings, like table already exists, the sequence already exists, and some privileges
couldnot be set. After those warnings it ran for more than 1/2 hour to load the table. The problem is that after this
wasall done. I went back into the target instance, and checked the newly loaded table. There was nothing in it, i.e., 0
rows.
>>
>> The pg_dump and the output file claim the pg_dump version was 12.6, while the server it ran against was 11.9. Both
instancesare AWS RDS Aurora-PostgreSQL ones. Any ideas why there was no data in the table? Other than on inputting it
usingpsql, there were no other errors.
 
>>
>> --jay
>>
>>
>>
>
>

-- 
Angular momentum makes the world go 'round.