on 11-16-2018 7:26 AM
How do i load LONG datatypes in BODS? I have tried loading data keeping LONG as datatype but, the job runs forever. I tried removing these long data type columns and running the job, it ran quickly. I have heard long_to_varchar also causes performance issue. Any suggestions?
Well, you might've already figured it out by yourself, but I stumbled over the very same problem (DB2 --> BODS(4.2) --> MS-SQL-Server) and was truly frightened by the "performance" of my dataflow. And I'd like to share my solution.
My Task was to load ~245K rows with 3 columns ( BIGINT, VARCHAR(100) but always NULL, CLOB(1048576))
I created the default DF (Source-Query-Target) and was punished by >20 minutes execution time.
I came to realize, that - try as I might - I cannot circumvent the "set-commitsize-to-1-because-of-long". That simply works as designed. Altering the source-fetch-size did not have any impact at all. But here's what I could do: Altering the target-options!
1. Set "Target-Commitsize" to 1 (just to avoid the log-entry)
2. Set "Number of loaders" to 20
3. DF: Source --> Target
The above settings resulted in ~5 minutes execution time, 🙂
I expect these settings to vary in accordance with your network-speed. For example, raising the # of loaders above 20 or the source-fetch-size to 2K didn't gain me anything.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Hi Karthik,
With LONG both read and write will be done row-by-row. See if you can avoid using LONG.
KBA 2311856 - Slow performance processing LONG datatypes - SAP Data Services
KBA 1430342 - Long type causing dataflow to be slow - Data Services
Thanks.
Akhilesh
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
What are you doing with your data? Writing them out to another table? That may be the reason for the lack of speed. With long colums the commit size is set to 1.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
User | Count |
---|---|
84 | |
10 | |
10 | |
9 | |
7 | |
6 | |
5 | |
5 | |
4 | |
4 |
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.