Re: [dabo-users] PostgreSQL char fields are rpadded

On 11/18/11 10:27 AM, Ed Leafe wrote:

> On Nov 18, 2011, at 12:18 PM, Paul McNett wrote:

>> Ooh, here's an inconsistency, since in Python "abc " != "abc". Since the database

>> does the padding/unpadding, the char values should come to the Dabo layer unpadded.

> If you want that behavior, make it varchar, not char.

Yes, that's what I'm doing now.

> What if you stored "abc " in a char(20) field? What would be a consistent value to receive back?

The inconconsistency that I'm presently talking about is that when comparing fields

in postgres, as with "where <field_name> = <other_field_name>", the padded characters

are insignificant. However, when comparing the strings in Python the padded chars are

significant.

Since postgres turns "abc" into "abc " when updating, and since postgres strips

the padded spaces in comparisons, it seems to follow that when psycopg2 fetches the

data from postgres, it should strip out the trailing spaces in char fields so that

python comparisons of the data don't take into account the padded spaces.

Paul

_______________________________________________

Post Messages to: Dabo-users@leafe.com