compare column with time stamps with +- tolerance

[ralfiii]

Hello!

I have files that come in the following format:
timestamp with 0.1 sec accuracy <Tab> Type

e.g.
Time Type
00:00:00.5 2
00:00:01.3 2
00:00:02.1 2
...

I'd like to be able to compare 2 such files.
*) If types differ that is a strong difference.
*) If the type is identical but the time difference is < 0.3 sec that is less important.

Is that achievable with BC?
(Naturally I can write converters that convert the timestamp e.g. into a float value if that helps.)

[ralfiii]

re-post...

Uups. Just found an old posting, it seems I already had the same problem a couple of years ago...

http://www.scootersoftware.com/vbulletin/showthread.php?t=4543

[Aaron]

How did the previous suggestion work for you? Defining the milliseconds into their own column.

[ralfiii]

Defining milliseconds into a separate column doesn't work as times like

00:00:01.9 and
00:00:02.1

or

00:00:59.9 and
00:01:00.1

would be reported as different even though they are only 0.2 sec apart.

Now I wrote a converter to automatically re-format files the following way:
*) 1st column = a sort of line number
*) 2nd column = time in seconds
*) 3rd col = time in "normal" format

BeatNr RelTime Time Type
0 1,2 00:00:01.2 2
1 1,9 00:00:01.9 2
2 2,6 00:00:02.6 2
3 3,4 00:00:03.4 2

Then I marked the 3rd column as unimportand and specified a numeric tolerance of 0.3 for the 2nd column. That works perfectly (BC, the best program EVER

But: I have to make this session setting changes every time I want to compare such files. Or I make that the default session then ALL csv-comparisons use that session settings.

Is it possible to link this session settings to a specific file name extension (e.g. "*.bea") - so that "normal" .csv file remain the way they are and .bea files use this configuration?

Thanks in advance!

[ralfiii]

resolved!

Forget my request.
I need to write a module that supports the validation anyway, so comparing with BC - which is still the best software ever - is not needed anymore.

Ralf