Re: Repeat: how do I erase mistakes from ~/.idata?

Skip Montanaro <[email protected]> writes:

>     python ifile cron cars
>     9000 1204 6975 6240
>     28 6 24 18
>
> and I delete "cron" as a category.  Then the first three lines will be:
>
>     python ifile cars
>     9000 1204 6240
>     28 6 18
>
> All frequency pairs that begin with "2:" will be deleted.  All frequency
> pairs that begin with "3:" will now begin with "2:".  Any words that no
> longer have any associated frequency pairs - because they were only ever
> associated with the category "cron" - will be deleted altogether.

i maybe wrong here, but in the light of this it seems that only the
aging is affected.  that's the number right after each word, the
second column.  so the static stats would be right, but maybe the odd
word is in the database a bit longer if it was "refreshed" in a
later removed category.

but what the heck:  the alternative to using your script were to
delete every message using -d, which is impossible unless you really
keep every message.  as a quick hack it propably does its job.  have
you (or anybody else) noticed any irregularities at all?

clemens