Why group by satement crash?

[#1]

The script is simple:

Load

Cust_no,

OCC_Date,

sum(AMT1) as AMT1,

sum(AMT2) as AMT2,

sum(AMT3) as AMT3

From AMT.qvd(qvd)

Group by Cust_no, OCC_Date;

If records of AMT.qv is small,such as 100,000 . but if the record is larger such as 100,000,000, qv will crash?

How to solve this problem?

[#2]

hi

your script is right please send me screen short of error you are facing.

Regards

Ashish

[vgutkovsky]

Well, for 100,000,000 rows you need a pretty powerful server, and you certainly can't do it on a desktop version. What are your server specs? Monitor memory/CPU usage and see what happens when it hits that group-by.

Regards,

[#4]

In Qv9SR5, My Server memory is 32GB. It seems not memory crash.

Why?

[vgutkovsky]

Try first loading the data to a table and then doing the GROUP BY from the resident table (instead of from the QVD directly). Also, is this a regular LOAD or a JOIN LOAD? I've seen problems like that arise with extremely large joins.

Regards,

[suniljain]

32 GB is too enough for 10 crore of record.