Thread: binary tipes

binary tipes

From
Mª José Sempere
Date:
Hello

We're searching information about binary type fields.
We've found two possible types: bytea and blob.

We've created a table with a bytea type field.
We'd like having no more than one row with identical content
In that field. So, we've declared a unique index on that field.
We'd like to know if in tables with a lot of rows the existence of that
index will slow the insertion of new rows (perhaps because before inserting
it must compare that field with the content of the same bytea field
For all the existing rows in the table).

We'd also like to know how does PostgreSql implement that comparation
Between bytea fields (if it compares all the content of the field,
Or if that comparation is improved with any algorithm, as CRC,
Or anything else).


If anybody has experience in that kind of problems, we'd
thank very much its help.

Greetings for everybody




Re: binary tipes

From
Peter Eisentraut
Date:
Am Freitag, 3. Dezember 2004 13:18 schrieb Mª José Sempere:
> We've created a table with a bytea type field.
> We'd like having no more than one row with identical content
> In that field. So, we've declared a unique index on that field.
> We'd like to know if in tables with a lot of rows the existence of that
> index will slow the insertion of new rows

Yes.

> (perhaps because before inserting
> it must compare that field with the content of the same bytea field
> For all the existing rows in the table).

Not with all rows, since it's a B-tree index, but with some rows for sure.

> We'd also like to know how does PostgreSql implement that comparation
> Between bytea fields (if it compares all the content of the field,

Yes.

--
Peter Eisentraut
http://developer.postgresql.org/~petere/