Re: Simple table creation - Mailing list pgsql-novice

Hi Chetan,

Thank you for replying.

I realise this is not a regular use case, but here is my situation.

I am taking over a reporting role in a company that is currently using a COBOL based ERP system.  Due to the nature of the system, data extract is very slow, so overnight extracts are used for day to day reporting purposes,  these extracts are stored as dbase files.

Other dbase files containing meta data have also been created to aid reporting.

This situation has led to many dbase files being used with many mashup combinations created for different report use cases, this is proving to be quite a maintenace overhead.

My preference is to dump all of the raw data into Postgress where I can create the needed indexes and views needed for all of the reports,  then the only data that needs updating is the source data, not a myriad of dbase files.

Unfortunately the raw data consists of many tables with greater than 100 columns, given the time constraints I am faced with I cannott commit the time needed to create the tables in postgress manually.

I have python scripts that I am gradually adopting to incrementaly replicate the data once the structure is in place.

I have had some limited results using LibreOffice Base.  I create a database connected to my postgress database and another connected to the COBOL ODBC connetor, then drag and drop tables from one to the other.  There are problems with the detection of some data types, and once corrected it is very slow and raises an error at about 4 Million rows leaving the data in an unknown state

I would love to find a way to do this without error, it would save me so much time.

Hope you can help.

Once I have the current situation under control I plan to introduce the business to the Kimball approach.

Cheers

Craig