Search Results

We all know what must be done.

The issue is how to do it.

Please, provide your solution using actual unix commands.

Here is one way of doing it:
#!/bin/ksh
sed 's/.*\(__.*__.*\).xml/\1/' Inp_File | uniq | while read mStr; do
grep "${mStr}" Inp_File | sort -t_ -k3 -kn1 | tail -1
done

See if this works for you:
#!/bin/ksh
mMax=$(sort -t, -n -k3 b | tail -1 | sed 's/.*,//')
egrep ",${mMax}$" b

See if the ^ (caret) works for you:
/au/äu/
/a[^u]/ä/
/o/ö/
/[^a]u/ü/

Please, display the desired output.

#!/bin/ksh
mX="https://abc.com/dir1/dir2/dir3/dir4/somefile.java"
echo ${mX} | sed 's#.*dir3/##'

If the user selects a grouping field, would a grep be sufficient?

Can you give us some examples of what the user could select?

If every file has always two records that you do not want to count,
then multiply the number of files by two and then subtract this number from the total wc.

To find the number of files:
ls -1 |...

Not true - you are not trying at all.

I gave you a complete solution in this thread and you continue to ask for alternative solutions.

What do you really want to accomplish with so many...

Can you provide a sample of the file?

If you had noticed, I made the two file names as variables for the shell.

This way, to have your desired output, you just run it twice:
1)

mF1='1.txt'
mF2='2.txt'

2)

mF1='2.txt'...

You had a very similar post that I provided a solution using shell scripts:
https://www.unix.com/shell-programming-scripting/176428-grab-unique-record-different-files-condition.html

Change it...

Here is a skeleton for one possible solution:
#!/bin/ksh
IFS='|'
while read mBefore mAfter; do
if [[ "${mBefore}" != "${mAfter}" ]]; then
echo ${mBefore} | sed 's/,/|/g' | read mB1 mB2 mB3...

Here is a solution using shell scripts:
#!/bin/ksh
typeset -i mFromA mToA mFromB mToB
mF1='1.txt'
mF2='2.txt'
mPrevTag=''
#### sort is used to reduce the number of "grep"
sort ${mF1} | while...

I used:
GNU bash, version 3.2.0(1)-release (x86_64-unknown-linux-gnu)

As much as possible, if one solution can be given in a more universal usage, then we must provide it.

Otherwise, for the...

You should have displayed the desired output.

Thus, making easier to understand your requirement.

In any event, an easy way to find if records of file A is present if file B:
grep -f FileB...

Ran above code using bash and received the following error:
grep: /dev/fd/62: No such file or directory

The following code works:
array1=(1234567 7665456 998889 000909)
array2=(1234567 5581445...

Please show us what you have done so far.

You should have the logic and ask for help on which commands to use.

See if this works for you:
#!/usr/bin/ksh
typeset -i mCnt
typeset -i mFlag
rm -f File.out
while read mLine; do
mCnt=$(echo ${mLine} | grep -c 'finally[ ]*{')
if [[ ${mCnt} != 0 ]]; then
...

Based on the input, can you display the desired output?

For your situation, you must use "&&" (and).

See if this works for you:
sed 's/\([A-Z0-9]\),/\1|/g' File

Extracted from the 'sed' book:

You can use:
strings -n1 File

This is one way of doing it:
sed 's/Search/Replace/' Input > Output