Re: Parse each line by character location

[Lie]

On Nov 5, 2:29*pm, Dennis Lee Bieber <[e-mail]:
[snip]

* * * * So you have a classic (especially for COBOL and older FORTRAN) fixed
field record layout, no?

* * * * I presume the entire file is of a single layout? That would mean
only one splitting format is needed...

[snip]

>
* * * * Note that all fields are still in character format. And has been
noted, I'm sure, if you try to turn the third field into an integer, you
may have problems, depending upon how you do the conversion -- leading 0
implies octal if it were a literal, though it seems int() handles it
correctly (Python 2.5)

from help(int)
| ... If base is zero, the proper base is guessed based on the
| string content. ...

int(x) will always convert x in base 10
int(x, 0) will convert based on literal int notation, i.e. the prefix
'0x' is base 16 (hex), prefix '0' is base 8 (octal), everything else
is base 10.
int(x, n) will convert on base n, where 2 <= n <= 36

if you're still in doubt, just pass the base explicitly to the int:
int(x, 10) (superfluous though)

[Shawn Milochik]

I work with tab-delimited files for a living. Because of the same need
you have, I created a Python script to do this. It has usage
information that is easy to follow (just run it without any
arguments).

I hope someone else finds this useful. I have, and use it every month.
It can be easily modified to create comma-delimited files, but that's
something I never use, so it does tabs.

http://milochik.com/shawn/fwconvert.zip

Usage:
fwconvert -r rulesFile fileName [-t|-f]
or
cat filename | fwconvert -r rulesFile" (-t|-f)

-t (to tab) or -f (to fixed-width) required when piping input to
script. Otherwise, it will be auto-determined.
Rules file format:
fieldStart:fieldLength,fieldStart:fieldLength...
Example:
1:3,4:20,24:5