Script to Find Lookups to Files

[Lerebours, Jose]

Brian posted:

> 
> > Since I've gotta task of renaming filePro files to get rid 
> of dots in the
> > names, I recall there once was a script that someone 
> provided here that
> > would show you all of the lookups that a particular file 
> made, or had made
> > to it (?).
> >
> > This kinda script would make my task sooooooo much easier 
> to do.  Was it
> > Stockler's Utilities thingamabob?
> 
> I and probably others have probably posted little one or 
> two-liner shell 
> commands to do it. There is more than one way.
> 
> Here is a quick-n-dirty, but it'll just fail if the total 
> number of files is 
> too large.
> It looks for all lines with lookup and . in the then section.
> 
> cd /u/appl/filepro
> egrep -n '.*:.*:lookup .*\..*' */prc.* |less -S
> 
> This handles unlimited number of files:
> cd /u/appl/filepro
> find ./ -name 'prc.*' |xargs egrep -n '.*:.*:lookup .*\..*' |less -S
> 
> or |pg if you don't have less, but this less syntax will not 
> wrap the long 
> lines, making it a little easier to read
> 
> Here is an awk-based way that does the same as above.
> 
> cd /u/appl/filepro
> find ./ -name 'prc.*' |xargs -n1 awk -F: '{if($3 ~ "lookup" 
> && $3 ~ ".") 
> print ARGV[1]": "NR": "$3}' |less -S
> 
> the output is:
> filename: line number:"then" section of the line
> 
> The advantage here is that you can add code to the awk script 
> to get more 
> and more accurate filtering and even perform actions on a per 
> line and per 
> file basis.
> For example you could add some more if() statements to 
> recognize and ignore 
> comments, recognize and ignore browse formats, recognize 
> lookups where the 
> target is a variable and include them rather than ignore them 
> so you can go 
> look at the processing and figure out if those variables 
> would contain a dot 
> at run-time yourself.
> It could even write out an edited copy of the process table where the 
> lookups have all been changed from dots to underscores and 
> even run system() 
> with B. Stocklers cabe script to compile the new tok.
> It can get more and more fancy depending on if it's worth the 
> bother, which 
> depends on just how many files you have and just how many 
> substitutions you 
> have to do.
> The more files and the more broken lookups, the more it's 
> worth getting a 
> sophisticated script to do more of the work automatically.
> 
> Think of awk script like rreport processing. It has 
> equivalents for @once, 
> @done, a system() command, and the main body runs once for 
> each record, 
> etc...
> 

Simple tools one should have in his toolbox.  These simple scripts
(or commands) are things of beauty.

Regards;


Jose Lerebours