Houssem Hariz
Houssem Hariz

Reputation: 55

EXPORT CSV with hana client hdbsql

I face some performance issues with exporting table as CSV to a remote server with HANA Client.

Before the Update to REV 07 the performance was acceptable but since one month we need 1 hour to export 7MB.

I searched in Internet and i find an archived question answered by Lars Breddemann.

https://archive.sap.com/discussions/thread/3748102

Lars says: HDBSQL is not particularly slow, but will of course suffer from network bandwidth and latency limitations, while the server side HANA export simply avoids that completely.

I didn't really inderstand what it means and I would like to know if someone know how to configure this latency or network bandwidth or maybe if there any Option to add in my command line to get better result.

Here my Command line

hdbsql -n XXXX:30015 -i 00 -u XXXXX -p XXXXX 1 -x -resultencoding LATIN1 -a -o "....\g5.csv" "SELECT * FROM SCHEMA.TABLE"

Thanks in advance and best regards

Houssem

Upvotes: 1

Views: 8065

Answers (1)

Lars Br.
Lars Br.

Reputation: 10396

What I meant by my comment was, that hdbsql just like any other client tool communicates with the HANA server via network. This involves delays, such as latency, which are due to very nature of how signal transfer works. The actual latency and the bandwidth usually can only be influenced by changing how the network connection is set up. For example, if you have a dial-in connection with a modem or a satellite connection latency (delay) typically will be higher than with a direct LAN connection. It's not something you can configure with 'hdbsql'.

SAP HANA provides an option to export table data as CSV on the server, without the need for a client tool. Whether this is a possible solution for your scenario needs to be checked.

However, you wrote that you had been happy with the performance as long as you were using SPS 07 (which is several years outdated by now) and now this is not the case anymore. What has all changed here? Which software version are you using now?

In order to find out more about the situation, you may use the SQLDBC trace option of hdbsql (-T <tracefile_name>). You may also want to check whether the communication to the HANA server actually is the problem. Maybe writing the output file is slowing the process down.

Upvotes: 1

Related Questions