UNIX for Dummies Questions & Answers

Hi,
This is my first time on this forum..
I searched the previous answers, but didn't find the answer I was looking for at first glance.

csplit works beautifully for me, except for one thing. My file looks like this:

ad|name1|asdf...(several pages)..asdf ...
ad|name2|asdf...(several pages)..asdf ...
ad|name3|asdf...(several pages)..asdf ...
...x500 ...

where 'name x' is some regular string

With the command:
csplit file.txt '/ad/' {500}

I was able to split the file into 500 files, each starting with "ad|name1|..."


My only problem is that I want the NAMES of these files to be whatever's in the correspoinding vertical bars. In the above example, for file1, I want the name to be name1.

I have no clue how I would go about naming the children files according to what's in between the vertical bars [ I already know that the vertical bars will NOT show up anywhere else in the file ]. Is there a way using csplit ??
---------------
If csplit does not have this capability, I guess I'll have to just rename the 500 files. Is there a way to rename files based on what string is found between specified characters ??

I hope I don't have to do this manually.

Thank you everyone in advance! Any help is MUCH appreciated

csplit's a new on on me so if someone knows a commandline arg for that that does it - sing out!

To do the rename afterwards though, you can do the following:

Code:

for file in ad\|*; do mv "$file" `echo "$file" | cut -d '|' -f 2`.txt; done

This is assuming the files really have | symbols in the names (if not, post a few of the filenames generated).
I've tested this but not exhaustively, so try it someplace safe of cource

Thank you very much Smiling Dragon for your response.

How do I apply that command ??

I just apply it like any command in UNIX ?? (like for example 'mkdir name' ) ??

I tried that exact command and got:
"mv: cannot stat `ad|*': No such file or directory

Yes, that's a regular one-liner command.
Have you confirmed that the filenames are as per my stated assumption?

That 'cannot stat' error is telling you that mv can't find any files beginning with ad|
I'm guessing that means the filenames are not quite as I've assumed, if not, please supply the exact format of them and clarify what you are after

What shell are you running this in? (find out by running "ps")

Oh I'm so sorry! I missed everything below your quoted portion!
* I'm still new to this forum! *

I didn't realize you had particular file names,
The filenames generated by csplit are in the format:
xx00
xx01
xx02
...
xx99
xx100
xx101
...
xx500

I don't think they have any extension, but if they do, it would be .vi

Thanks again!

Oh I see, you want the filename to be changed to match the part between the | characters within the file itself?

If so:

Code:

for file in xx*; do mv "$file" `head -1 "$file" | cut -d '|' -f 2`.txt; done

Should do the trick.

I've picked a .txt extension for the new filename, but you can change it whatever you wish.

Note that this will run on anything starting with xx so be careful what else is in that directory when you run this.

Better yet, test it someplace safe first.

Thank you so much Smiling Dragon!

I finally got a chance to try this.
I just realized that I made a mistake in the question I originally asked.

This code you suggested:

for file in xx*; do mv "$file" `head -1 "$file" | cut -d '|' -f 2`.txt; done
works perfectly for what I asked for, but I just noticed that my files are actually in the format:

asdf...asdf | some Junk | asdf...asdf |name1| asdf... .... and no more vertical bars

I was actually supposed to name the files after what is between the SECOND set of vertical bars, rather than the first.

Is there a way to do that ??

And by any chance, now that I've renamed 500 files incorrectly, is there an undo button in unix ??