Datadeploying attachment of size 50mb to Oracle Blob.

Have a dct which has a attachment field with replicant size 20. When a pdf of size 2.4Mb is attached to each of those 20 replicants and when deployed to oracle DB takes about 10-13 mins.

Oracle table has id (number) and Image(Blob) cloumn each attachment is deployed in a seperate row and the Datadeploy takes close to 10 mins.

Where in a OD those new 20 pdfs (different ones ) just takes less than a minute to deploy to the file server.

Any reason why such a huge difference in performance?

Find more posts tagged with

Comments

Bill Klish

DataDeploy is creating temporary tables behind the scenes to move those large chunks of data around. The performance has always been notoriously slow when we have used clob or blob data in the past, but it has gotten better in more recent releases. What version of OD/DD are you using?

You may want to play around with the JVM memory settings for the DataDeploy process as well. That has yielded better performance in the past.

amigo

Srini,

I have a similar requirement in my project for deploying uploaded files to a blob coumn in the DB. The attach files field is not a replicant.
Currently I am using a cgi callout for attaching the files.
cgi-callout url="/iw-bin/promptUpload.ipl" label="Upload File" window-features="width=320, height=200, resizable=yes, toolbar=no, scrollbars=yes"/
But when I datadeploy this field to a blob column, only the filename seems to be inserted in the column....not the entire file object as such.
It would be really helpful if you could provide any pointers/sample code on how to deal with this.

Thanks in advance.

Adam Stoller

Srini,

I have a similar requirement in my project for deploying uploaded files to a blob coumn in the DB. The attach files field is not a replicant.
Currently I am using a cgi callout for attaching the files.
cgi-callout url="/iw-bin/promptUpload.ipl" label="Upload File" window-features="width=320, height=200, resizable=yes, toolbar=no, scrollbars=yes"/
But when I datadeploy this field to a blob column, only the filename seems to be inserted in the column....not the entire file object as such.
It would be really helpful if you could provide any pointers/sample code on how to deal with this.

Thanks in advance.

Usually callouts in DCTs only store the path/filename in the DCR - so that's all that would be deployed to the DB. If you want to change such references to become actual contents of the files then you either need to modify your callout (to store the contents of the file into a field - UGH - prone to limitations of the size of a textarea) - or perhaps you can use some sort of tuple pre-processor (IMO they tend to be complex and difficult to write/debug) - or (my preference) use something like Perl DBI instead of DD and that way you can code the logic directly.