All Forums Teradata Studio
ZackHowe 15 posts Joined 07/06
22 Oct 2015
Teradata Studio Bug

Using Studio 15.10.4
if I issue select * from very_big_table it will think a while and then return no rows.  with a success code.
if I select * from very_big_table sample 2147483646 I get the records.  If I go one more, up to the max 4 byte int limit...no rows, no error message. 
Using old-school sql assistanct the query works too.

fgrimmer 553 posts Joined 04/09
22 Oct 2015

Zack, I don't think you can attach your error logs here so please open an incident with Support center and include your logs. Thanks.

ZackHowe 15 posts Joined 07/06
22 Oct 2015

Francine,
 
The problem is that there is no error.  It shows that the query completes successfully, and no rows were returned. 
but in fact there are 3+ billion rows in the table.
It shows the run time...a couple minutes, and in viewpoint I can see it spooling the rows for return.  But then nothing....
It should be easy to reproduce, just select * from a table with more than 2.1 billion rows :)
 
I also tried it with just selecting a single column, the PK (a bigint) of the table to ensure that it wasn't some strange data character, or an artifact of "too much data" instead of too many rows.  but the same failure with 1 column at 2.1 billion rows...

fgrimmer 553 posts Joined 04/09
22 Oct 2015

Zack, I suspect we are hitting a row count limit in JDBC. We'll try to recreate it. thanks.

ZackHowe 15 posts Joined 07/06
26 Oct 2015

Great.  If I can be of some kind of help, let me know.  but I hope you can reproduce it easily enough :)
 

MikeDempsey 94 posts Joined 10/06
26 Oct 2015

I had the same issue in SQLA many years ago. I used an internal 32 bit counter.
The solution was to remove the counter since I did not have access to a 64bit integer at that time. Now I would have simply changed the counter to a 64 bit integer.

You must sign in to leave a comment.