Shell Programming and Scripting

Hello

I have a file with list of numbers

Code:

380634062
380907172
380733031
380733032
380896961
380896962
381067621
381067622
380995932

I would like to find a way to fill in a sed command delete pattren

Code:

sed -i '/$number_from_list/d' -e '/$second_from_list/d' -e ... etc

so I can end up with a long sed command which contains all search pattern and run it once against the file.


I had my script setup to use for loops to loop through the content of my text file and do delete everytime it takes a number from the list. but this makes it very very slow! I'm trying to run all numbers against the file and delete all lines which starts with those numbers once and for all.

Thank you

Moderator's Comments:

Please use next time code tags for your code and data

Trying to use sed for this makes it much harder than it needs to be.

Code:

grep -f excludefile inputfile > outputfile

grep -vf , no?

Yes, good point. -v to reverse the output to exclude the matches instead of including them.

oh yeah I did that on my initial run to do invert stuff!

the thing is its running! but still very slow! is there is a faster way?

The grep version is extremely fast. Please show exactly how you're using it, there must be other problems with your script.

If the files are enormous(hundreds of megs and up), the limit may be your disk speed.

Alternatively try:

Code:

awk 'NR==FNR{A[$1]=1;next}!A[$1]' excludefile inputfile > outputfile