[SOLVED] How to check if a keyword is in a part of file?

massy · 11-27-2019, 01:36 AM

I have a file with a pattern as like as below:

...
Start block
32
56
4545
89
23
End block
...
kdshf
ksdfdk
89
jdfh
...

I need to check if 89 is exist just in the block(start to end). I don't want to find 89 after the block.
How can I do this by sed or grep or awk in a script?
--------
New request:
Indeed I need to change the line: commentting some of items just in the block. My ideal output is a thing as below:
...
Start block
32
#56
#4545
#89
23
End block
...
kdshf
ksdfdk
56
4545
89

jdfh
...

Turbocapitalist · 11-27-2019, 02:15 AM

There are several ways to get there.

Code:

sed -n -e '/^Start/,/^End/{/^89$/p}' massy-02.data

test 89 -eq $(sed -n -e '/^Start/,/^End/{/^89$/p}' file.data ) && echo OK || echo Not

berndbausch · 11-27-2019, 07:28 AM

Quote:

Originally Posted by massy

Please tell me whenever I'm wrong in English writing...

I am not a native speaker myself, so take this with a grain of salt.

Quote:

I have a file with a pattern as like as below:

with a pattern as below. No need for as like, and as like doesn't make much sense at this point anyway.

Quote:

I need to check if 89 is exist just in the block(start to end).

... if 89 exists ...

Quote:

How can I do this by sed or grep or awk in a script?

I would use with instead of by, but I am not certain if by is incorrect here.

Regarding the technical answer, you could use a more complicated awk program, but Turbocapitalist's solution is probably the most succinct one.

massy · 11-27-2019, 08:16 AM

Quote:

Originally Posted by Turbocapitalist

There are several ways to get there.

Code:

sed -n -e '/^Start/,/^End/{/^89$/p}' massy-02.data

test 89 -eq $(sed -n -e '/^Start/,/^End/{/^89$/p}' file.data ) && echo OK || echo Not

Now I need to change the line: commentting some of items just in the block. My ideal output is a thing as below:
...
Start block
32
#56
#4545
#89
23
End block
...
kdshf
ksdfdk
56
4545
89

jdfh
...

massy · 11-27-2019, 08:22 AM

Quote:

Originally Posted by berndbausch

I am not a native speaker myself, so take this with a grain of salt.

with a pattern as below. No need for as like, and as like doesn't make much sense at this point anyway.

... if 89 exists ...

I would use with instead of by, but I am not certain if by is incorrect here.

Regarding the technical answer, you could use a more complicated awk program, but Turbocapitalist's solution is probably the most succinct one.

Thanks a lot :-).

MadeInGermany · 11-30-2019, 06:56 AM

Written as a multi-liner

Code:

sed '/^Start block/,/^End block/{
# we are within the block
# if the line is 89 then put a # at the beginning
/^89$/ s/^/#/
# if the line is 4545 then ...
/^4545$/ s/^/#/
}' input_file

With -i option you can write back to the input_file:

Code:

sed -i.bak ...

The (optional) .bak leaves a input_file.bak behind.