Deleting lines from a file with binary pattern strings - bash

Question

I've got two files. File A contains text written in N lines, and File B contains a binary pattern string of 0 and 1 that has N length too.
I want to delete the lines from File A that has the same line number that the one on File B that contains a 0.
I've read that it might be a good idea to do it with awk, but I don't have any idea of how to use it.
Files are very long, like 2000 lines for example (they are video traces)
For example:
File A:
Line 1: 123456
Line 2: 789012
Line 3: 345678
Line 4: 901234
File B:
Line 1: 1
Line 2: 0
Line 3: 0
Line 4: 1
After the execution:
File A:
Line 1: 123456
Line 2: 901234

Answer 1

You can use paste and cut for this:
paste fileB fileA | grep '^1' | cut -f2-
paste fileB fileA - pastes file contents side by side, delimited by a tab
grep '^1' - filters that lines that start with 1
cut -f2- - extracts the content that we need
Both cut and paste use tab as the default delimiter.
This is very similar to Benjamin's solution. A small advantage here is that it would work even if fileA were to have more than one field per line.

Answer 2

Assuming Line 1: etc don't really exist in your input files all you need is:
awk 'NR==FNR{a[NR]=$0;next} a[FNR]' fileB fileA

Answer 3

You could use a decorate – filter – undecorate pattern:
paste fileA fileB | grep -v '0$' | cut -f1
This prints the lines of each file next to each other (paste), then filters the lines that end with 0 (grep), then removes the lines from the second file (cut).
This breaks if fileA contains the delimiter used for paste and cut (a tab by default). To avoid that, we could either swap the files (see codeforester's answer) or resort to something like
paste fileA fileB | sed -n '/1$/s/\t.$//p'
(if line ends with 1, remove tab and last character, then print) or
paste fileA fileB | grep -Po '.*(?=\t1$)'
(match only lines ending in 1, use zero-width look-ahead to exclude tab and 1 from match); the last solution requires a grep that supports Perl compatible regular expressions (PCRE) such as GNU grep.

Answer 4

Lots of interesting answers here. Here's a bash one:
while IFS= read -r -u3 line; IFS= read -r -u4 bool; do
((bool == 1)) && printf "%s\n" "$line"
done 3<fileA 4<fileB
This will be much slower than other solutions.

Answer 5

another paste/awk solution. If tab appears in data find another delimiter.
paste file2 file1 | awk -F'\t' '$1{print $2}'

Answer 6

A single awk command can read from both files.
awk '(getline flag < "fileB") > 0 && flag' fileA
After reading each line from fileA, read a line from fileB into a variable flag and test if its integer value is true or not. For true values, the line from fileA is printed.
Depending on your version of awk, you may need to use int(flag) or flag+0 to force the value to be treated as an integer rather than an ordinary non-empty string.

Answer 7

EDIT: #codeforester's comment if Line 1 or Line 2 are not part of your File1 and File2 then following may help.
awk 'FNR==NR{a[FNR]=$0;next} $0!=0{print a[FNR]}' filea fileb
Solution 2nd: Reading fileb file first and then reading filea then.
awk 'FNR==NR{if($0!=0){a[FNR]=$0};next} a[FNR]' fileb filea
Solution 1st's alternative in case OP has string(s) line1, line2 in his/her files.
Following awk may help here too.
awk '
FNR==NR{
a[FNR]=$NF;
next}
$NF!=0{
printf("%s%s\n","Line " ++count": ",a[FNR])
}' filea fileb

Answer 8

paste and sed combo:
paste -d'\n' fileB fileA | sed -n '/^1$/{n;p}'
123456
901234
You interleave the files:
1
123456
0
789012
0
345678
1
901234
Then you use sed to print the lines that follow directly a line that has only a 1. However this will not behave properly if you have entries that are composed only of a 1 in the fileA. If it is the case then you have to use the following sed command that takes into account if we are currently processing an odd/even line:
paste -d'\n' fileB fileA | sed -n '1~2{/^1$/{n;p}}'