print on single line after comparing the no. and GREP

[RVSachin]

Is there a easy way to do the following in (Korn) shell?

My text file contains lines similar to the following, each on a different line:

	#  reg1.inst_abc[0].a unmatched
	#  reg1.inst_abc[1].a unmatched
	....
	....
	.....
	#  reg1.inst_abc[n].a unmatched

	# inst_abc[0].a unmatched
	# inst_abc[1].a unmatched
	...
	...
	...
	# inst_abc[n].a unmatched

[highlight]Note that n can be any no. and need not be in a sequence.[/highlight]


I want these lines sorted out and [highlight]printed on a single line [/highlight]such that I print the following lines into a file:
reg1.inst_abc[0].a matches with inst_abc[0].a
The no. 'n' on left should match the one on the right side.

It will be of great help if the commands used to achieve this is explained.

Thanks,

RV

 

[RVSachin]

Can someone help me, please!

 

[Mike042]

Hi there,
I don't think there is an "easy way to do the following in (Korn) shell?", otherwise someone would have posted it by now. There are a number of questions that need answers, before a complete solution can be found, eg:
Will there always be matching pairs of entries ?
What happens if there are not ?
What is the range of values for the no. 'n' ?
Although the numbers can be out of sequence, are there any missing? and is that an error if there are ?

So without making too many assumtions, how about the following ideas,

First idea:
- cut the source file in half by grep'ing into 2 files the different lines (reg1.inst_abc[n].a into one file and inst_abc[n].a into the other)
- sort the 2 files (beware that if you use a character sort, then when n is 21 to 29 those lines will appear between the lines when n is 2 and n is 3)
- read the files, or paste the files (process them in some way) to produce the desired output
- check that all lines from the source file are in the output file, any non-shows would be an error.

Second idea:
- grep the file for an incrementing count in square brackets (eg grep -F [${count}] <file>) in a while loop (till count greater than the no. 'n')
- check that 2 lines are present and output the combination of these 2 lines in the desired format to the output file
- as a final check, test that the important bit of each line of the source file (ie reg1.inst_abc[n].a and inst_abc[n].a) are present in the output file, any non-shows would be an error.
This idea automatically sorts the lines into true numeric order. This fulfills your requirement "I want these lines sorted out and printed on a single line"

Third idea:
- process the source file line by line (read while do loop) and save the important part of the line (ie reg1.inst_abc[n].a) in an array variable (eg array[n])
- when the second item is read (ie inst_abc[n].a ), extract the no. 'n', get the contents of array[n] and output the two items in the desired format to the output file and empty the array variable
- at the end, any array variables that contain values must be errors so print them in a list.

Three ideas for you to peruse and develop further into a script to do the job.

I hope that helps.

Mike