compare two files and delete records matching in the first file

[chapm4]

Think awk may be the solution but not sure.

Example:

File 1 File 2
M02464217A M03234567A
M01234567A M02464217A
M02345678A
M03456789A

If data in File 2 matches data in File 1, I want to remove matching data from File 1. In above example I would remove M03234567A from File 1.

 

[marsd]

This works, but it's ugly.

awk ' {
while ((getline I < &quot;target&quot > )) {
if ($0 == I) {
print $0 >> targetfile
next
}
}
}'

 

[CaKiwi]

Hi chapm4,

I assume you want M02464217A deleted from the first file. If not you can change this program to do what you want. Run it by entering

awk -f awkscript file1 > file3

BEGIN {
  while ((getline < &quot;file2&quot;) > 0) {
    hld[++ix] = $0
  }
}
{
  flg = 0
  for (i=1; i<=ix && !flg; i++) {
    if ($0 == hld[i]) flg = 1
  }
  if (!flg) print
}

Hope this helps. CaKiwi

 

[marsd]

typo:
> 0 not > )

Problem:
you still need to write to a temp file and remove the original this way.
tried also:
for xx in $(cat file)
do
if grep $xx secondfile
then
sed &quot; s/$xx// &quot; file >> holdfile
fi
done

 

[CaKiwi]

or you could change it in place by modifying marsd's slolution to use ex

for xx in `cat file2`
do
  echo "g/$xx/d|w" | ex - file1
done

CaKiwi

 

[aigles]

If input files are sorted (it's not your case) you can use the join command.

join -v1 file1 file2 > file3

In this simple example, join compare the first field of lines of each files and output unpairables lines in file1. Jean Pierre.

 

[aigles]

Another version of CaKiwi solution :

awk -f awkscript file1 > file3

BEGIN {
while ((getline < &quot;file2&quot > 0) {
hld[$0] = 1 ;
}
}
{
if (! hld[$0]) print $0
}

Jean Pierre.