how to find (grep) paragraph start with word and ending with other word and ignore new line?

Question

I want to find a paragraph is a list of files. For example an SQL statement which looks like:

upadte b_table
set number = 100014
where id in (
            select number_desc from t_table
            where id > 100);

How should I ignore enter/new lines?

I need to find/print a paragraph which starts with "update" and finishes with the first occurrence of ;

Answer 1

Inspired by barryj's concise awk solution:

perl -ne 'print if /update/../;/' *.sql

If the update statements are contained in (empty-line delimited) paragraphs:

perl -00 -ne 'print if /update.*?;/' *.sql

-- from http://uselessuseofcat.com/?p=381

Otherwise, this prints filenames and update statements:

perl -ne 'BEGIN{undef $/}; print "$ARGV\t$.\t$1\n" if m/(update.*?;)/mg' *.sql

-- from http://www.commandlinefu.com/commands/view/5087/multi-line-grep

(untested - caveat emptor)

Answer 2

grep is a bit tricky with newlines, but you could try something with awk:

awk '/update/,/;/' < filename.sql

Answer 3

I used grep like this:

grep -ihRP  '^(update|( )* update) .* [^;]*;'