string manipulation in ksh 1

[SimonSellick]

Hi,

I am faced with filenames of a format x-CCYYMMDD.csv where x can be anything between 1 and 14 characters, examples:

T001-20050901.csv
XY_AB_12_01-20031231.csv

I need to extract the date string and the preceding code into shell variables. For the two files above, the results should be:

$FILENAME $DATE $CODE
T001-20050901.csv 20050901 T001
XY_AB_12_01-20031231.csv 20031231 XY_AB_12_01

Is there a substring() or equivalent function available in the shell, or in awk; or can anyone suggest another way to achieve this? I can rely on the hyphen separator between code and date, but I can't rely on its being the only one (that is, there might be some in the code section as well).

Any pointers gratefully received.

 

[PHV]

FILENAME="XY_AB_12_01-20031231.csv"
CODE=${FILENAME%-*}
DATE="${FILENAME#*-}";DATE="${DATE%.csv}"
echo "$FILENAME\t$DATE\t$CODE"

Hope This Helps, PH.
Want to get great answers to your Tek-Tips questions? Have a look at FAQ219-2884 or FAQ181-2886

 

[SimonSellick]

Yes, that's great - thanks PHV.

Now all I need to do is work out why it works!

 

[feherke]

Hi

I can rely on the hyphen separator between code and date, but I can't rely on its being the only one

FILENAME="with-more-dashes-20031231.csv"
eval `echo "$FILENAME" | sed 's/\(.*\)-\([^-]*\)/CODE="\1";DATE="\2"/'`
echo -e "$FILENAME\t$DATE\t$CODE"

Feherke.

http://rootshell.be/~feherke/

 

[PHV]

OK, replace this:
DATE="${FILENAME#*-}";DATE="${DATE%.csv}"
By this:
DATE="${FILENAME##*-}";DATE="${DATE%.csv}"

Hope This Helps, PH.
Want to get great answers to your Tek-Tips questions? Have a look at FAQ219-2884 or FAQ181-2886

 

[SimonSellick]

OK, got it (found the ksh man page once you'd shown me what to look for). And the ## matches the longest possible string before the terminating hyphen, rather than just up to the first one.

Clever stuff.

Thanks again both.