cancel
Showing results for 
Search instead for 
Did you mean: 
Read only

CSV-Import taking forever

fabian_krger
Participant
0 Likes
587

Hello,

I used to import small .csv Files (~120 KB, 3000 Entries) into one of my tables (HANA dev edition on cloudshare).

Until yesterday, everything worked out, the import took a few seconds (nut sure, but something like 10, maximum 20 seconds).

Yesterday I loaded 3 files and after that I could not load any further file.

IMPORT FROM CSV FILE '/tmp/98.csv'

INTO "BenchmarkValues"

WITH

          THREADS 1

          BATCH 10000

          COLUMN LIST IN FIRST ROW

          FIELD DELIMITED BY ','

          ERROR LOG '/tmp/BenchmarkValues_98.log';

then results in:

The CPU remains at 100% (1 core, but the cores change over time) and it does not end. I waited over 40 minutes but it seems to have no end.

I can see the Process at "Administration --> Performance --> Threads" and it has the status:

The error log is empty:

Seems like cancel the thread does not work (it just continues), I have to restart the instance to stop this.

By using selects I can see that actually no data is inserted (at least not commited). After the restart there is also nothing of new data inserted.

Here is what I tried so far:

- Switching auto merge off for this table --> no effect

- Setting the Batch size to 1000 (a number smaller than the file entries) --> no effect

- Manually inserting the first entry via insert statement --> works well

- Check if there is enough free disk space --> more than 100 GB free?

- Reclaim log --> no effect

What else could I check or try? I am running Rev 48...

Any hints appreciated!

Kind Regards,

Fabian

View Entire Topic
Former Member
0 Likes

Have you tried to load manually from Hana Studio?

File Menu -> Import -> SAP Hana Content -> Data from Local file...

I Faced similar problem, it was due to file format (Junk characters at end of file).

-Jayant

fabian_krger
Participant
0 Likes

Thanks for the hint.

I looked a little bit closer at the files.

I created them by saving an .xls Table as .csv.

All of them are ANSI as UTF-8 but the last ones (those who don't import) are Macintosh instead of DOS\Windows.

I will try to convert this and see if it actually solves the problem...


EDIT: Thanks, I was now able to solve the problem. Don't know why but had many other errors with the Line delimiters. Got it finally working by specifying RECORD DELIMITED BY '\r\n' ...