extract part of the file/read next lines if condition is fullfiled on the current/awk

[pet_ra]

Hi,All!

Need one's help.
i have a big file looking like this

<page>
<title>...</title>
<..../>
....
</page>...

that is what i want to extract:

<page>
 <title>Talk:Atlas Shrugged</title>
    <id>128</id>
    <revision>
      <id>152717854</id>
      <timestamp>2007-08-21T16:32:33Z</timestamp>
      <contributor>
        <username>Marlith</username>
        <id>4871029</id>
      </contributor>
      <minor />
      <comment>/* The quotes do not belong here */</comment>
      <text
xml:space="preserve">{{NovelsWikiProject|class=B|importance=High}}
......
</text>
</revision>
</page>

so what i need is

if('<title>Talk:')

-> start printing/store in other file up to

if(!'<page>')

will appear, then stop reading and search for

if('<title>Talk:')

again

I don't get how can i achieve to read lines after

<title>Talk:

- sed and grep are reading each line and next or getline is not enough for me. Is there any possibility to define global flag or any other ideas?

Please if possible comments with examples.

Thank you in forehand,
Zina

[radoulov]

awk '/<page/ { tag = $0 }
/<title>Talk:/ { $0 = tag "\n" $0; f=1 }
f && /<\/page/ { print; f=0 } 
f' filename

[pet_ra]

Wow! it works perfectly! Thank you very much!
Anyway i don't want to left dummy. Let me understand the way of your coding.

you find <page and save it in tag
next you read line and it is not <page but it submits the next condition <title then you edit currently read line by adding tag "\n" $0; and define a flag f=1
next you read the third line and first and second condition are not fulfilled but flag is 1 and </page condition is not full filled - the line is $1, then going on with reading and comparing with all three conditions(<page, <title, </page) - lines are $2-${infinity}

after we met

</page

we print everything starting from $0 and set our flag to zero. Now we look again for fulfilling the condition.

Is it so?
But really wow!
Thank you and respect,
Zina

[radoulov]

Wow! it works perfectly! Thank you very much!
Anyway i don't want to left dummy. Let me understand the way of your coding.

you find <page and save it in tag
next you read line and it is not <page but it submits the next condition <title then you edit currently read line by adding tag "\n" $0; and define a flag f=1
[...]

So,
when the current record matches the pattern "<title>Talk:",
we modify the current record $0: we prepend tag (the previous record)
and a new line, then we set f (our flag) to 1.
We continue, if the flag is set to 1(true) AND the current record matches
the pattern "</page>", we print it and set the flag to 0(false).
The final f do the real work: it reads all the records and if the flag is set to 1(true), it prints the record.