Is there a way to find middle of a pattern and lowercase it in bash

Question

I am working on decreasing the size of text data.

Example input:

[email protected];example
[email protected]:exmaple

Example output:

[email protected];example
[email protected]:exmaple

Pseudo code:

if line has "@" and ":" or ";"
replace the text between @ and : or ; with lowercase

But I have no idea even what tools to use. Any help is welcome.

Any of these helpful?: https://stackoverflow.com/questions/2264428/how-to-convert-a-string-to-lower-case-in-bash — James Brown, May 10 '19 at 12:58
Is there some reason you aren't lowercasing all of it? Email addresses are not case sensitive, so why not standardize the whole thing? — Paul Hodges, May 10 '19 at 13:02
Unfortunately the documents contains passwords after the : or ; — John W, May 10 '19 at 13:05

score 0 · Answer 1 · edited Jun 20 '20 at 09:12

Use the sed tool for resolve this problem,

cat input_file.txt | sed -e 's/\(.*@\)\([A-Za-z.]\+\)\([;:].*\)/\1\L\2\3/'

Regex Explanation:

\(.*@\) - This pattern matching "example@"

\([A-Za-z.]\+\) - This pattern matching "EXAMPLE.com"

\([;:].*\) - This pattern matching ":exmaple" or ";exmaple"

The \L is changed to lowercase of the text

If you want to update the content use -i flag in sed command.

Ex:

sed -i -e 's/\(.*@\)\([A-Za-z.]\+\)\([;:].*\)/\1\L\2\3/' input_file.txt

Paul Hodges · Answer 2 · 2019-05-10T13:40:28.430

If you have a lot of data, awk will be faster than a shell. The sed solutions are fine, but this works too:

$: awk '-F[;:]' '{ printf "%s;%s\n", tolower($1), $2 }' x
[email protected];exaMple
[email protected];eXmaple
[email protected];exAmple
[email protected];exmaplE
example_example.com;Example
example_example.com;eXmaple
[email protected],example;

That defines the -Field separators as a list of ;: and lowercases the first field. I arbitrarily replaced the delimiter with a standardized ; - if that doesn't work, this might not be the best solution for you. Stick with the sed.

sprabhakaran beat me to it with a practically identical sed solution while I was initially typing, lol. :)

sed can.

$: cat x
[email protected];exaMple
[email protected]:eXmaple
[email protected];example
[email protected]:exmaple
example_EXAMPLE.com;example
example_EXAMPLE.com:exmaple
[email protected],example

$: sed -E '/@.+[;:]/s/^(.*)@(.*)([;:])(.*)/\1@\L\2\E\3\4/' x
[email protected];exaMple
[email protected]:eXmaple
[email protected];exAmple
[email protected]:exmaplE
example_EXAMPLE.com;Example
example_EXAMPLE.com:eXmaple
[email protected],examPle

\L says to begin lowercasing until \E (end) or \U (begin uppercasing).

This skips lines that don't have both @ and [;:] (either of ; or :.)

for small datasets native bash might be easier.

It might be a lot simpler however to just downcase the whole thing.

$: declare -l line
$: while read line
> do echo "$line"
> done < x
[email protected];example
[email protected]:exmaple
[email protected];example
[email protected]:exmaple
example_example.com;example
example_example.com:exmaple
[email protected],example

declare -l make a variable always lowercase anything put in it.

Since case-sensitive passwords prevent that, parse the parts separately.

$: while IFS="$IFS:;" read email pass
> do echo "$email [$pass]"
> done < x
[email protected] [exaMple]
[email protected] [eXmaple]
[email protected] [exAmple]
[email protected] [exmaplE]
example_example.com [Example]
example_example.com [eXmaple]
[email protected],example []

As long as the record is properly formatted it works great. I assume you can check for errors or trust your data.

Is there a way to find middle of a pattern and lowercase it in bash

2 Answers2

Since case-sensitive passwords prevent that, parse the parts separately.