UNIX for Dummies Questions & Answers

Hi

I know very little about unix.
I have a unix code (which I'm running in geektools) which gives me the word count of a pages document. However, the pages document innately contains 522 words of superfluous "Ipsem Lorem..." text which are included in the count.
How would I go about subtracting 522 from the word count expressed by my code, which goes something like:

gunzip path/filename.pages | cat index.xml | xsltproc path/index.xsl | wc -w

Thank you!

"grep -v " can exclude the line with "Ipsem Lorem"

Code:

grep -v "Ipsem Lorem" YOUR_FILE |wc -w

lorem ipsum surely?

Anyway there appears to be a problem with your pipeline.

If we assume that you are processing a file called "index.xml" from the archive filename.pages, we need to remove one of the pipe characters. Then the pipeline makes sense. We can then adjust the total (assuming that the faulty pipeline was not giving you wrong figures by reading "index.xml" before gunzip finished).
My guess at the pipeline may be flawed - see below.

Code:

gunzip path/filename.pages
words_total=`cat index.xml | xsltproc path/index.xsl | wc -w`
if [ ${words_total} -ge 522 ]
then
	words_total=`expr ${words_total} - 522`
fi
echo $words_total

Afterthought: The whole script could be flawed. I am not familiar with xsltproc and its parameters and don't know if each archive includes a file called "index.xml" or where "path/index.xsl" comes from. Can you try each program in turn to be sure that the flow (particularly to pipelines) is not flawed and that the output of the pipeline is something which can be counted with "wc -w". I'm starting to wonder whether this should be a sequence of commands with no pipeline at all?

Other then the fact that it won't update unless i change the code and change it back (still working on that) the code seems to work fine
The index.xsl is a script file i was told to download and save. Index.xml is within the .pages zip bundle. The whole thing outputs a wordcount.

Thanks for the code. After trying I realized that the wordcount was still off. I tried adding and subtract different numbers. It was a hundred too high, so I subtracted 422. Then it was about 500 to low, so I added a hundred. The numbers keep changing. It's really mysterious, but I've got the word count close enough for my purpose (that is if I can get it to update D

You may find it beneficial to break the task down into separate components and then test each one individually. I experienced the variable results when trying your pipeline approach. I still wonder whether it should be a sequence of consecutive commands with the last command piped to "wc".