need help parsing

AutoC · 08-03-2009, 12:55 AM

Hi,
I have a file with entries as follows..

Quote:

|
|
point [ p oy n t ]
pointed [ p oy n ] [ t ah d ]
|
|

I need to rewrite them as

Quote:

|
|
point p_oy_n_t
pointed p_oy_n t_ah_d
|
|

what can I do for this?
any help is appreciated

Meson · 08-03-2009, 12:58 AM

You need to be a little more general, that looks really specific. How do you know what spaces get replaced by underscores and what spaces stay as they were? syllables?

Can you generate a longer example?

AutoC · 08-03-2009, 02:47 AM

1. the word remains as it is.
2. each set of characters enclosed in braces are merged using underscore
eg. [ a bc d ] becomes a_bc_d
that is, the left brace and space between left brace and first character is eliminated.similarly, the right brace and space before it.
all spaces inbetween are made underscore.

3. one space between each of the groups enclosed by braces.
eg. [a bc d ] [e fg h] becomes a_bc_d e_fg_h

here's a long example

according [ ah ] [ k ao r ] [ d ih ] [ ng ]
has to become
according ah(one space)k_ao_r(one space)d_ih(one space)ng

ghostdog74 · 08-03-2009, 03:39 AM

Code:

awk -F"]" '/\[/{
    for(i=1;i<=NF;i++){
        match($i,/\[/)
        o=substr($i,1,RSTART)
        s=substr($i,RSTART+1)
        gsub(/ +/,"_",s)
        gsub(/^_|_$/," ",s)
        $i = o s       
    }
}1' OFS="]" file

output

Code:

# more file
|
|
point [ p oy n t ]
pointed [ p oy n ] [ t ah d ]
|
|
# ./test.sh
|
|
point [ p_oy_n_t ]
pointed [ p_oy_n ] [ t_ah_d ]
|
|

AutoC · 08-03-2009, 03:54 AM

awesome.thanks

Kenhelm · 08-03-2009, 05:48 PM

Using GNU sed

Code:

echo '
point [ p oy n t ]
pointed [ p oy n ] [ t ah d ]
according [ ah ] [ k ao r ] [ d ih ] [ ng ]' |

sed -r ':a s/(\[[^]]*) /\1_/; ta; s/\[_|_\]//g'

point p_oy_n_t
pointed p_oy_n t_ah_d
according ah k_ao_r d_ih ng