[SOLVED] How to add characters and lines to file

nobtiba · 01-29-2011, 11:32 PM

Dear Bros,

I have this issue but haven't figure out the solution:
For example, I have a file called "file" like this one:
type=strongsubj len=1 word=absolve pos=verb stemmed=y priorpolarity=positive
type=strongsubj len=1 word=unique pos=adj stemmed=n priorpolarity=neutral
type=strongsubj len=1 word=absolutely pos=adj stemmed=n priorpolarity=neutral
type=weaksubj len=1 word=taking pos=verb stemmed=y priorpolarity=positive
type=weaksubj len=1 word=friend pos=noun stemmed=n priorpolarity=positive
type=weaksubj len=1 word=usually pos=adverb stemmed=n priorpolarity=positive
type=strongsubj len=1 word=purecolor pos=anypos stemmed=n priorpolarity=negative
type=strongsubj len=1 word=accusingly pos=anypos stemmed=n priorpolarity=negative

I want to add the plural for the noun, for example if find this line:
type=weaksubj len=1 word=friend pos=noun stemmed=n priorpolarity=positive
will add one more line :
type=weaksubj len=1 word=friends pos=noun stemmed=n priorpolarity=positive
where we add "s" for the word friend
I did try to do like that:
<code>
cat file | while read LINE ; do

set -- ${line}
if [[ "${4#pos1=}" == "noun" ]];then
#I tried this line but it doesn't work properly:
v3==$(echo $line |sed 's/$3/$s') #I want to find the third word "word=friend" in that line and add "s" after that word
# I don't know what command to add this new line "$v3" to the file ???

done
</code>

Could you please give me some hint ?

paulsm4 · 01-30-2011, 12:02 AM

Hi -

There are two questions here:
1. How do you parse the input line (e.g. in order to intelligently find "friend" and change it to "friends")?

2. How do you add the new line to the end of the file?

It sounds like you've actually figured out "1)", the hard part. You implied you're getting a valid string in "$v3". If so, cool! If not, I'd urge you to consider exploring "awk". Or writing the whole thing in Perl or Python.

The easy part is "2)":

Code:

echo $v3 >> myfile

dugan · 01-30-2011, 12:04 AM

Here's a Python solution:

Code:

word_index = 2
pos_index = 3
output = []
with open("file") as f:
    for line in f:
        output.append(line)
        tokens = line.split()
        pos = tokens[pos_index].split('=')[1]
        if pos == 'noun' and tokens[word_index][-1] != 's':
            tokens[word_index] = tokens[word_index] + 's'
            output.append(' '.join(tokens))

for line in output:
    print line

MALCOMEX999 · 01-30-2011, 02:40 AM

With awk...

Code:

awk '{split($4,noun,"=");if(noun[2]="noun"){$3=$3"s"}}1' infile

grail · 01-30-2011, 02:59 AM

Again with awk:

Code:

awk -F"[ =]" '$8 == "noun"{$6=$6"s"}1' file

nobtiba · 01-30-2011, 10:22 PM

Quote:

Originally Posted by grail

Again with awk:

Code:

awk -F"[ =]" '$8 == "noun"{$6=$6"s"}1' file

@MALCOMEX999: I tried your solution but it seems to add "s" for every word.
@grail: your solution look very short and work... I dont really understand the "awk" command so I dont know how to add this to my script? (or it work alone ?)
This:
=====================================================
cat file | while read LINE ; do
set -- ${line}
awk -F"[ =]" '$8 == "noun"{$6=$6"s"}1' file
done
=====================================================
give me infinite loops, never stop.
What I want is after double the line it found (find the noun then copy a new line below it) then modify it (add "s"). Finish. When we open that file, we should able to see the new line with addition.
Will this work ?

grail · 01-31-2011, 01:06 AM

No you don't need the bash while loop and so on as it will stand alone.
Easiest way to do what you are asking is:

Code:

awk -F"[ =]" '$8 == "noun"{line=$0;$6=$6"s";$0=line"\n"$0}1' file

nobtiba · 02-01-2011, 07:16 AM

Quote:

Originally Posted by grail

No you don't need the bash while loop and so on as it will stand alone.
Easiest way to do what you are asking is:

Code:

awk -F"[ =]" '$8 == "noun"{line=$0;$6=$6"s";$0=line"\n"$0}1' file

Your solution give the right result but it only print to the screen, not write to the file it self, so it will be hard for me to copy only the line with "s" to the file manually.
Is there any way to solve ?

nobtiba · 02-10-2011, 08:49 PM

Quote:

Originally Posted by grail

No you don't need the bash while loop and so on as it will stand alone.
Easiest way to do what you are asking is:

Code:

awk -F"[ =]" '$8 == "noun"{line=$0;$6=$6"s";$0=line"\n"$0}1' file

Hi grail,
its my fault not to see the result carefully. Yes, it add one line with modification as expected !
Thats great job !
Can this command do more than one modification ? For example, not only add "s" after the "noun" but also add "s" after verb, "ed" after verb.
for example if we have this:
type=weaksubj len=1 word1=abandonment pos1=noun stemmed1=n priorpolarity=negative
type=weaksubj len=1 word1=abandoned pos1=adj stemmed1=n priorpolarity=negative
type=weaksubj len=1 word1=abandon pos1=verb stemmed1=y priorpolarity=negative
type=strongsubj len=1 word1=abase pos1=verb stemmed1=y priorpolarity=negative

it will becomes this:
type=weaksubj len=1 word1=abandonment pos1=noun stemmed1=n priorpolarity=negative
type=weaksubj len=1 word1=abandonments pos1=noun stemmed1=n priorpolarity=negative
type=weaksubj len=1 word1=abandoned pos1=adj stemmed1=n priorpolarity=negative
type=weaksubj len=1 word1=abandon pos1=verb stemmed1=y priorpolarity=negative
type=weaksubj len=1 word1=abandons pos1=verb stemmed1=y priorpolarity=negative
type=weaksubj len=1 word1=abandoned pos1=verb stemmed1=y priorpolarity=negative
type=strongsubj len=1 word1=abase pos1=verb stemmed1=y priorpolarity=negative
type=strongsubj len=1 word1=abases pos1=verb stemmed1=y priorpolarity=negative
type=strongsubj len=1 word1=abased pos1=verb stemmed1=y priorpolarity=negative

Which is quite similar job but how to make it do the job one time, not repeat and make redundancy?

grail · 02-10-2011, 09:35 PM

Yes just follow the process. You can either combine all changes to a 'verb' or place adding 's' in part and 'ed' in another.

nobtiba · 02-10-2011, 11:38 PM

Quote:

Originally Posted by grail

Yes just follow the process. You can either combine all changes to a 'verb' or place adding 's' in part and 'ed' in another.

I tried this:
awk -F"[ =]" '$8 == "noun"{line=$0;$6=$6"s";$0=line"\n"$0}1;$8 == "verb"{line=$0;$6=$6"s";$0=line"\n"$0}1{line=$0;$6=$6"ed";$0=line"\n"$0}1' file

but it works differently, not as expected... May be I understand the parameters wrongly ?

grail · 02-11-2011, 12:18 AM

Your first issue is the '1' after each section. The '1' is there to tell awk to print (ie the default action for a true expression).
The second is that any pair of braces {} not proceeded by an expression (ie $8 == "verb") will run for every line in the file.
So I am guessing your example, apart from printing multiple entries due to first point, is adding 'ed' to every line irrelevant of being a 'verb'.

nobtiba · 02-13-2011, 08:40 PM

Thanks Grail ! Great help !