[Solved] Database linked csv, problem with field types

[Hagar Delest]

Hi there.

I have a base file linked to a directory of csv files.

My problem is, I can't seem to change how the fields are read in the database.

I've read somewhere that they are decided on by reading the second line of the csv (after the headers) and then whatever datatype is present in the column will decide the field type for the whole column.

I've done some testing with this and it looks like I'm getting random results with how the data is read. I want the data as a number type so I can make a SUM with the column when imported to calc. however quite often the data is read as a TEXT field and causes problems.

the csv file is updated daily from the internet hence the csv directory structure.

Any help would be appreciated.

[RoryOF]

A small sample file would help.

[Villeroy]

The text driver for Base has not enough import options to meet common requirements. It works fine with unquoted numbers and all numerals according to the global setting. You can set the decimal separator independently from any locale but "special numbers" (mostly dates, times, booleans)" are interpreted in the global locale context.

[blah_blah_blah]

Thanks for the help,

I ended up going with what is explained in
https://forum.openoffice.org/en/forum/v ... 83&t=23260