Changing file name with sed command

Question

[jordi.planells]

Hi all. I need a little bit of help. I have a list of files in a folder, they look like this:

   H3K4me1/Cell-Line=S2-DRSC#Developmental-Stage=Late-Embryonic-stage#Tissue=Embryo-derived-cell-line/ChIP-chip/Rep-1//Dmel_r5.32/modENCODE_304.gff3

H3K4me2/Cell-Line=S2-DRSC#Developmental-Stage=Late-Embryonic-stage#Tissue=Embryo-derived-cell-line/ChIP-chip/Rep-1//Dmel_r5.32/modENCODE_965.gff3

H3K4me3/Cell-Line=S2-DRSC#Developmental-Stage=Late-Embryonic-stage#Tissue=Embryo-derived-cell-line/ChIP-chip/Rep-1//Dmel_r5.32/modENCODE_3761.gff3

H3K9ac/Cell-Line=S2-DRSC#Developmental-Stage=Late-Embryonic-stage#Tissue=Embryo-derived-cell-line/ChIP-chip/Rep-1//Dmel_r5.32/modENCODE_3765.gff3

I would like to change their name with sed command, so the final name would be like this:

H3K4me1_modENCODE_304.gff3

H3K4me2_modENCODE_965.gff3

H3K4me3_modENCODE_3761.gff3

H3K9ac_modENCODE_3765.gff3

I know i can play with sed s and the / but I don't manage to get it to work. I would extremely appreciate if you explain a bit the code you give as an answer.

Thank you in advance! Best wishes,

Jordi

[Pierre Lindenbaum]

every time you put a '#' or a '=' in a filename, god kills a kitten.

[jordi.planells]

got them named like this from modENCODE page, but thanks for the information, never will dare to do it (poor kittens!)

[Joe]

Just be a little cautious with all the /'s in those names. That'll look like a directory structure to bash.

e.g. touch 'H3K4me1/Cell-Line=S2-DRSC#Developmental-Stage=Late-Embryonic-stage#Tissue=Embryo-derived-cell-line/ChIP-chip/Rep-1//Dmel_r5.32/modENCODE_304.gff3' fails miserably, even when hard quoted.

[dariober]

+1 Pierre

every time you put a '#' or a '=' in a filename, god kills a kitten.

Then we have the culprit for wars and famines: spaces in filenames.

Answer 1

[Joe]

This is all you need in terms of an expression:

sed 's|/.*/|_|gi'

e.g.

$ echo "H3K4me1/Cell-Line=S2-DRSC#Developmental-Stage=Late-Embryonic-stage#Tissue=Embryo-derived-cell-line/ChIP-chip/Rep-1//Dmel_r5.32/modENCODE_304.gff3" | sed 's|/.*/|_|gi'
H3K4me1_modENCODE_304.gff3

Pro tip, use that regular expression with the rename program.

Explained:

s = substitute

| = sed delimiter so as to avoid confusion with the more traditional /

/ = Match the first forward slash you find

.* = followed by any character, any number of times

/ = until you meet another /.

|_| = replace all the previously matched stuff with an underscore.

gi = globally, and case insensitive (you don't strictly need these, I include them as force of habit).

[jordi.planells]

Thank you! Your answer is very much appreciated

[Joe]

Glad it helped. Be sure to accept it to provide closure to the thread if it resolved your problem.