Looking to take an input file that

[stemp1ar]

Looking to take an input file that is not in the proper format and
change the file to the proper format.

current format is: 1234-7654321 Characters - more chars,1234

I am able to do what I need in multiple iterations of a file, but what
I really would like is to change the file in one pass.

requested format: 1234|7654321|Characters|1234|today's date|more characters

currently have:
echo ' {print substr($1,1,4) "|" substr($1,6,7) "|"}
' > ./awkscript.txt

awk -f ./awkscript.txt ./datafile > datafile.new

looking for a regular expression using awk or any other suggestions that will work...

I am able to parse the first 2 fields, but I am very new to awk and having a bit of trouble with the syntax of reg expression pattern matching...I would think using a reg expression getting the $2 up the the comma delimiter would be the next step...any help would be great

Thanks

 

[CaKiwi]

Here is a solution using regular expressions in sed

sed "s/\(....\)-\(.......\)\([^-]*\)-\([^,]*\),\(.*\)/\1|\2|\3|\5|13Dec01|\4/"

You could also do it in awk using index to find the delimeter characters and substr to split up the line.
Hope this helps.
CaKiwi

 

[stemp1ar]

Thanks...I did use sed in my first iteration...

Is there a way in sed to use a variable in the replacement string?

sed "s/\(....\)-\(.......\)\([^-]*\)-\([^,]*\),\(.*\)/\1|\2|\3|\5|13Dec01|\4/"

such as: "s/\(....\)-\(.......\)\([^-]*\)-\([^,]*\),\(.*\)/\1|\2|\3|\5|$VARNAME|\4/"

In my attempts the $VARNAME prints exactly that...even attempting to to `$(date)` to execute the statement and suggestions???

 

[CaKiwi]

The only way I can think of doing it is to echo the sed program to a file with the date formatted appropriately in it

echo "s/ ... |`date %d%b%Y`| ... /" > sed.tmp

then run sed using this file as the program

sed -f sed.tmp ./datafile > ./datafile.new

I don't think there is a way to get sed to read its program from standard input.

Hope this helps. CaKiwi

 

[marsd]

The sed faq states that using the double quotes
as sed -e " s/pat/$VAR/g" text will work in this situation....

 

[stemp1ar]

Thanks, I just tried that before coming here and yes it does work in this situation...

Again, thanks for the help

 

[stemp1ar]

I have a question on the post above and the regular expression used:

sed "s/\(....\)-\(.......\)\([^-]*\)-\([^,]*\),\(.*\)/\1|\2|\3|\5|13Dec01|\4/"

The original input file: 1234-7654321 Characters - more chars,1234

In the example above, I am curious on the matching...

regexp used:/\(....\)-\(.......\)\([^-]*\)-\([^,]*\),\(.*\)/\1|\2|\3|\5|13Dec01|\4/

matching: /\(1234\)-\(7654321\)\ Characters - more chars,1234

I understand the first portion, but what is "([^-]*\)-\([^,]*\),\(.*\)" and )/\1|\2|\3|\5 and
\4/ ???

 

[CaKiwi]

stemplar,

A ^ as the first character in [] means that any character which is not in the [] will be matched i.e. [^-] matches any character which is not a -. [^-]* means match any number of characters which are not a -. The first set of characters matched in a \( \) pair is saved in register 1, the 2nd in register 2 etc. and they are replayed in the replacement string by specifying \1, \2 etc. I hope this is clear. Post again is you need further explanation. CaKiwi