How to print 10 letters preceding every occurrence of a particular character?

问题

Using grep, I can print all occurrences of the uppercase letter "Z" in my document. The output, however, will display the entire lines in which every "Z" in the document was found. I need to limit this to printing only the 10 letters appearing before every occurance of "Z".

E.g., if the document has a line with "AAAABBBBBBBBBCCCCCCDDDDDDDZ", it will print "CCDDDDDDD", the 10 letters appearing before.

If there are fewer than 10 letters prior to "Z", then nothing needs to be printed.
If "Z" appears multiple times in a single line, the 10 letters preceding each of these "Z"'s should be printed, e.g.: "AAAABBBBBBBBBZCCCCCDDDDDDDZ" will print "ABBBBBBBBB" and "CCDDDDDDDZ".

The result will be an output list of these letters, e.g.:

ABBBBBBBBB
CCDDDDDDDZ

How can I print the 10 letters preceding every occurrence of the letter "Z" in my document?

回答1:

Simple:

grep -oP '.{10}(?=Z)' <<< AAAABBBBBBBBBZCCCCCDDDDDDDZ

Explanation:

-o     : Print only match, not entire line
-P     : Use PCRE / Perl regex
.{10}  : Match is any 10 characters,
(?=z)  : which are followed by "Z". (Search for positive look-ahead for more details)
<<< ...: Here string

EDIT:

NOTE: This does not work, if the 10 characters we want are overlapping. e.g. input=AAAABBBBBBBBBZDDDDDDDZ. If the input contains such pattern, see igegami's answer

回答2:

$ perl -nE'say for /(?<=(.{10}))Z/g' <<'__EOI__'
AAAABBBBBBBBBZCCCCCDDDDDDDZ
AAAABBBBBBBBBZDDDDDDDZ
__EOI__
ABBBBBBBBB
CCCDDDDDDD
ABBBBBBBBB
BBZDDDDDDD

$ perl -nE'say for /(?=(.{10})Z)/g' <<'__EOI__'
AAAABBBBBBBBBZCCCCCDDDDDDDZ
AAAABBBBBBBBBZDDDDDDDZ
__EOI__
ABBBBBBBBB
CCCDDDDDDD
ABBBBBBBBB
BBZDDDDDDD

来源：https://stackoverflow.com/questions/17773501/how-to-print-10-letters-preceding-every-occurrence-of-a-particular-character

标签

perl

bash

grep