Gfilui

i have so many files under 300 different directories, file tree like:

--s1
 - ----abc
 - ----bcd
--s2
 - ----123
 - ----234

... etc.

I want to put them together under the same directory, like:

--whole
 - ----abc
 - ----bcd
 - ----123
 - ----234 ...

Is there any useful bash script in practical?

I coded this piece of bash script:

mkdir wavs

for ((i=1;i<=9;i++)); do
 cd ~/wav/train/S000$i
 mv * ~/wav/train/wavs
 cd .. done

for ((i=10;i<=99;i++)); do
 cd ~/wav/train/S00$i
 mv * ~/wav/train/wavs
 cd .. done

for ((i=100;i<=917;i++)); do
 cd ~/wav/train/S0$i
 mv * ~/wav/train/wavs
 cd .. done

echo "ok"

but i got the error which i dont understand:

./untar.sh: line 24: cd: /wav/train/S0917: No such file or directory 
cp: target '/wav/train/wavs/' is not a directory ok

Maybe the simplest way to do this job is by the command find (by default it works recursively):

find ~/wav/train/S* -type f -name "*.wav" -exec echo mv ~/wav/train/wavs/ ;

• ~/wav/train/S* is the search path and it will match to each sub dir that starts with S.




• -type f will limit the search only to the files.




• -name "*.wav" will limit the search only to the files that ends with .wav. Not mandatory.




• -exec ... ; will execute the mentioned command once for each search coincidence.




• is a variable that contains the coincidence item.




• remove echo from the command echo mv ~/wav/train/wavs/ to do the action.

Further if you want to delete the directories you can use a command as one of these:

find ~/wav/train/S* -type d -name "S*" -exec echo rm -r ;
find ~/wav/train/S* -type d -name "S*" -delete