Fred Williams wrote:
Would it not be more logical to modify the .import command to place a
null value in a column where the input value is non existent? (i.e. Back
to back separators in the input) Null in, null out.
Fred,
If you did that, there would be no way to enter an empty string as a
value. A NULL is not the same as an empty string.
Since the shell already defines a reasonable way to convert a NULL to a
sentinel string on output, it makes more sense to me to reuse that, than
it does create another special purpose mechanism to allow a user to
enter an empty string. If the shell recognizes the same sentinel value
on input, it can substitute a real NULL value.
The only danger I see is that the initial value of the .nullvalue
sentinel string may be an empty string (since it converts NULLs to empty
strings by default on output). In this case it would also convert empty
strings on input into NULLs, so we would have the same situation where
it's not possible to enter an empty string. To prevent this the
.nullvalue string should be initialized to a NULL and then no
substitution should be done on input unless the user changes it to some
other value. I don't think you can change the output handling (by
default a NULL and an empty string look the same). There would also need
to be a mechanism to set the .nullvalue string back to NULL to turn off
the input substitution.
Dennis Cote
-----------------------------------------------------------------------------
To unsubscribe, send email to [EMAIL PROTECTED]
-----------------------------------------------------------------------------
