CODEWITHSUNDEEP
linuxbashawk
Colin Moreno Burgess
Colin Moreno Burgess

Reputation: 1602

Search pattern bash variable to awk

Good morning

I've been developing a script to make certain search/compare/retrieve data on Linux systems. In particular what I'm trying to cross is data from 2 files previously created by the script with the following example text:

file_1:

/dev/abcde_vg/abcde_lv            APP     /bbdd_abcde                          
/dev/abcde_vg/system_lv           APP     /bbdd_abcde/system                   
/dev/abcde_vg/data_lv             DAT     /bbdd_abcde/data      

file_2:              

/bbdd_abcde         APP 
/bbdd_abcde/system      APP 
/bbdd_abcde/data            DAT

The goal here is to getting a list of the first column of the file 2 with awk:

for i in `cat file_2|awk '{print $1}'
do ..........;done

and to obtain something similar to the following statement:

awk '$3 ~ "^/bbdd_abcde$"{print $0}' file_1
/dev/abcde_vg/abcde_lv            APP     /bbdd_abcde

I've tried the following with no success:

for i in `cat file_2|awk '{print $2}'
do
awk -v VAR1="$i" '$3 ~ VAR1{print $0}' (or awk -v VAR1="$i" '$3 ~ "^VAR1$"{print $0}' or similars)
done

With no success at all.

Thanks for the answers, finally I applied a search for an exact match using AWK with this:

for i in $(cat file_2|awk '{print $2}')
  do
    awk -v VAR1="$i" '$3 == VAR1{print $0}'
  done

The idea behind this is more complex, the problem itself is part of a 200+ script for crossing data using bash only

The script generates a file on each server with a format like explained before

/dev/abcde_vg/abcde_lv            APP     /bbdd_abcde                          
/dev/abcde_vg/system_lv           APP     /bbdd_abcde/system                   
/dev/abcde_vg/data_lv             DAT     /bbdd_abcde/data

And those files are check against a "how it should be" file on a NAS with the format:

/bbdd_          APP 
/bbdd_/system       APP 
/bbdd_/data                 DAT

To verify if the "type" of disk assigned to an lv is correct. (disk types are figured out in another part of the script)

So the goal was to compare each mount point disk type "APP/DAT..." to the same in "how it should be"

the main problem for me was that searching for a specific mount point like "/bbdd_abcde" on one file with awk and "~" I got a lot of results instead of only the exact one. I needed to use "==" in my awk statement.

Thanks for the help

Upvotes: 1

Views: 306

Answers (2)

suspectus
suspectus

Reputation: 17288

This solution reads the cross-reference keys (APP, DATA) from file1 into an array then prints the matches with file2.

awk 'FNR==NR{a[$2]=$1;next} $2 in a' file2 file1

Upvotes: 1

Kent
Kent

Reputation: 195289

I hope I understood your question right, try this:

awk 'NR==FNR{a[$1]=7;next}a[$3]' file2 file1

note that, there is difference between this line and your codes. in my line, I didn't do regex match. because if you read /bbdd_abcde, with regex match checking all lines in your file1 would match. I don't know if it is expected.

If matching was desired, you can change the 2nd part of the codes.

Upvotes: 1

Related Questions