parsing column info

Running PostgreSQL 7.2.1 on RedHat Linux 2.4.7-10.

This might wind up being a perl hack, but I'll ask ...

I have a table that has been imported from someplace
(probably MS Access). The columns on the table has
been defined as character varying(50). One of the
columns is "supposed" to be defined as having a decimal.

I no longer want the decimals in that column. Is there
a way to parse out the decimals when copying this into
a new table? I was thinking that I can just create the
new table schema, select * from t_table and somewhere
figure out how to parse out the silly '.' from the column
(perhaps in a case statement?).

Any ideas?

Thanks!

-X