Question

Can you comment notes along with the code so I can follow along and understand what is being written?? Thanks in advance!

Let's say we're in a directory and we want to rename every file ending in·c, to end in .C (i.e., capitalize the c'). Here's one way to do it: $ for i in *. c; b-basename "$1" .c'; mv "8b.c" do "8b.C". done Things get more complicated if you want more complex renaming. Say for example you just created exactly one thousand simulation output files named output.XXX, where XXX runs from 000 through 999 inclusive. These files used a simulation parameter equal to 10, but you want to re-do the simulation changing that parameter to 20. So it would be useful now to rename all the files to look like output.p10.XXX, and then the new ones will be called output.p20.XXX. Then it gets harder. One way would be to do this: for i in output.*; do # hmmmmm, now what??? Maybe not a loop . is output. * l sed 's/output\.//, I awk ,(printf "mv output.%d output .p10.%dIn",$1,$1)' I sh # Yes, the above works. But also note you don't have to use ,/, to separate the source # from the target in sed. Whatever follows the initial ,s' is defined to be your # separator, so the above could just as easily have been : ,s; output\.;;, sed # use a semi-colon separator That last one provides a general hint: when algorithmically renaming files, regular expressions are far more powerful than shell globbing. So let's formalize that Create a program called rename. Its first argument is a sed substitution command, and al remaining arguments are existing filenames that w be renamed using the substitution command. For simplicity the user will leave out the initial 's', and the first character in the first argument shall be used as the separator Thus, both examples above can be done in one line each: rename '. p10.'output.* Note in the second case, only the first period will be matched. If the user wants to match every occurance of the source regular expression, the user can append a 'g' to the end of the substitution string, so for example "rename ;a;A;g' *" will change every lower-case 'a to an upper case one in all filenames in the current directory, To avoid catastrophe, rename should refuse to over-write an existing file, but it will only print a warning when that happens, and continue to perform the rename operation on all files for which the destination doesn't over-write an existing file. Finally, if given the "-f ("force") option before the substitution argument, then it will even over-write existing files with abandon. Thus, the man-page style SYNOPSIS is: rename [-f] sed-substitution-command file1 file2 Finally, to be maximally flexible, the files can actually be path names including slashes-which explains why we probably don't want to use ,「as the sed separator. So if the above examples were in some other directory, say "src" in the first case and "sim-output in the second, this should work fine: rename ';1.c$;.C; src/*.C rename '. p10.' sim-output/output.* It is OK to assume that only the last element of the path (ie, none of the intervening directories) should match the regular expression. I think this would be hard to verify on-the-fly, so you're allowed to just assume it

Answer 1

The question is to to rename .c file to .C extension in the directory

So there are various ways shown here, comments with respect to code are as:

1.

$ for i in *.c     // this line will iterate over all file which is ending with .c extension

do b = 'basename "$i" .c' ; //getting basename of file with extension .c

mv "$b.c" "$b.C"   //renaming .c files to .C extensions

done

2. Suppose you have complex renaming like output.XXX where XXX varies from 000 to 999 and earlier a simulation parameter was 10 so the file name were like output.p10.XXX, so now you want to change that simulation parameter to 20 then file name will be like output.p20.XXX. So first how to make files name like output.p10.XXX :

for i in output.*; //iterate over all the files in directory

do ls output.* (//give list of all the files with named output) | sed 's/output\.//' (//this will seperate output and XXX from file name, you can also use ; in place of / after s as a seperator of files given by ls but it can cause problem discussed later) | awk 'printf{"mv output.%d output.p10.%d\n",$1,$1}' (//this will make file output.XXX to output.p10.XXX and $1 is equal to XXX as using sed we seperated so the XXX is stored in 1 position and it is accessed through $1) | sh

3. Here you can write above thing in rename.sh file and execute directly as

rename ';\.c$;.C' *.c it will change all file .c to .C

rename ';\.;.p10.;' output.* it will change output.XXX to output.p10.XXX file name

4.

Here rename -f option is to force fully rename the files as without -f there will be warnings

5.

We should not use / in seperator as directory itself can have / in path name so that will cause problem as :

rename ';\.c$;.C' src/*.c it will change all file .c to .C

rename ';\.;.p10.;' sim-output/output.* it will change output.XXX to output.p10.XXX file name

src is different directory and sim-output is different directory. Here / is included in path so can't use / as seperator as discussed above.