Shell Programming and Scripting

Hi Friends,

I have a series of alphabets like this

Code:

[CA]A[GCAT][ATG][AA]GCAA[TG]

The values inside the square brace indicate that either of them can be present at that position. And those ones without a brace, means that they are the only ones that could be printed at that location.

Now, I would like to know if there is a possibility for a script to print all possible combinations of these alphabet occurrences.

For an example sake, if I consider only the first alphabet within a square brace, the output would be

Code:

CAGAAGCAAT

The alphabets in bold indicate the first ones in the square brace. Like this way, is there a possibility to print all possible combinations?

Thanks

Shell has property in which it expands anything given in braces:

Code:

$ echo AB{c,d}E
ABcE ABdE

So, its just a matter of replacing the square brackets with curly braces and putting comma between characters, and using the shell expansion property:

Code:

$ x='[CA]A[GCAT][ATG][AA]GCAA[TG]'
$ eval echo $(echo $x | perl -pe 'tr/][/}{/;s/\{([A-Z]+)\}/"{" . join (",",split("",$1)) . "}"/eg;')

Guru.

That was a great solution. Thanks guru! Instead of printing them side by side, is there a way to print one under the other.

Thanks once again for your time.

Brilliant solution indeed, guruprasadpr!

@jacobs.smith: try:

Code:

$ echo -e {C,A}A{G,C,A,T}{A,T,G}GCAA{T,G}"\n"

or

Code:

$ printf "%s\n" {C,A}A{G,C,A,T}{A,T,G}GCAA{T,G}

thanks for your comments.

@jacobs:
Modifying using printf as RudiC has used:

Code:

$ printf "%s\n" $(eval echo $(echo $x | perl -pe 's/\[([A-Z]+)\]/"{" . join (",",split("",$1)) . "}"/eg;'))

Also, tweaked the solution a little by removing the tr part.

Guru.

A variation on a theme which eliminates eval, echo, and the command substitution subshell:

Code:

perl -pe 's/.../.../eg; $_ = "printf %s\\\\n " . $_;' | sh

Obviously, sh should be a shell which supports brace expansion.

Regards,
Alister