command to print lines with occurence of a specific word more than occurence of a other word-CodePudding

For a text file:

A B A B A A
A A B A A
B B B A

expected O/P:
A B A B A A
A A B A A
{because [n(A) > n(B)]}

Note: It would be a great help, if the solution can be a command like grep or awk with condition explained above.

CodePudding user response：

You can do:

awk 'gsub(/A/,"&")>gsub(/B/,"&")' file 
A B A B A A
A A B A A

This works because gsub returns the number of substations made. The & in the replacement means replace with what was matched.

CodePudding user response：

This might work for you (GNU sed):

sed -E 'h;:a;s/A(.*)B|B(.*)A/\1\2/;ta;/A/!d;x' file

Make a copy of each line in the hold space.

Subtract A and B or B and A from the line until failure.

If an A remains, replace the current line with the copy and print, otherwise delete the current line.