removing punctuations from file input or standard input (UNIX)

lwj94 · 03-24-2014, 08:35 AM

I just began learning Unix and I'm trying to learn some stuff myself. How do I remove all punctuation from a file input or standard input? which means that the input must be called in a file and through a standard input method.

For example, if the input is:

"Hello", she said. "Nice to meet you!", he {thought aloud}.

Then the output should be:

Hello she said Nice to meet you he thought aloud

szboardstretcher · 03-24-2014, 08:43 AM

tr used with [ : p u n c t : ] (spaces added to lazily avoid smileys), would work in this case. Or you could use sed or awk.

A quick google of your question will turn up plenty of actual examples. Choose your poison.

e.g http://an-compsci.blogspot.com/2009/...-in-linux.html

danielbmartin · 03-24-2014, 09:02 AM

With this InFile ...

Code:

"Hello", she said. "Nice to meet you!", he {thought aloud}.

... this tr ...

Code:

tr -d "[:punct:]" <$InFile >$OutFile

... produced this OutFile ...

Code:

Hello she said Nice to meet you he thought aloud

The tr means translate.
The -d means delete.
The [:punct:] identifies the character class of all punctuation marks.

There are other ways to accomplish the same purpose.
I like this one because (IMO) it is more readable than the others.

Daniel B. Martin

pan64 · 03-24-2014, 09:07 AM

[noparse][:punct:][/noparse] can be used to avoid [

unct:]
sed 's/[[:punct:]]//g' works too...