On Sun, Feb 12, 2017 at 08:06:28PM -0500, bruce wrote: > Hey guys. > > Thanks for the delete replies.. > > Got a quick sed question now.. > > test file:: > head -2 sed.dat > 228d98f0_f16a_11e6_9544_1ad613f05f7b,1486934882 > 22b93712_f16a_11e6_a6ad_1ad613f05f7b,1486934883 > > want to simply truncate/search/replace the end of each line starting > with the "," > to get > > 228d98f0_f16a_11e6_9544_1ad613f05f7b > 22b93712_f16a_11e6_a6ad_1ad613f05f7b > > the following isn't working > sed -i 's/\,+\$//' sed.dat > > sed -i 's/\,+$//' sed.dat > > I'm sure it's simple/trivial... > Several problems. The sed's I know don't use extended RE syntax and '+' is a repetition character of extended RE. Some sed's might allow its use with the '-r' option. And maybe there are some sed's that do use extended RE. Comma is not typically a special character, so no need to escape. In pattern matching by the shell, the '*' repetition character does not reference the preceeding character. RE repetition characters (* in basic and + and ? in extended) always refer to repetitions of the preceeding character. Thus your pattern (if and extended RE) would be looking for a string of commas, NOT one comma followed by a string of any char. Try something like "sed -r -i 's/,[0-9]+$//'". or "sed -i 's/,[0-9]*$//'" or if uncertain they will be digits, but certain they will not be commas (i.e. you want to delete the last comma onward) try "sed -i 's/,[^,]*$//' Jon -- Jon H. LaBadie jonfu@xxxxxxxxxx _______________________________________________ users mailing list -- users@xxxxxxxxxxxxxxxxxxxxxxx To unsubscribe send an email to users-leave@xxxxxxxxxxxxxxxxxxxxxxx
