In a Bash script I want to split a string into two other strings based on the last "/" it contains.
In a situation where the given string is "Example/Folder/Structure", I would like to create two other strings with the following values:
string 1 = "Example/Folder"
string 2 = "Structure"
I'm trying to create a script to get a slather coverage report for a given folder in an iOS app. Although I have minimal knowledge of Bash, I was able to get it to work when the given folder is located in the root of the project. Now I want to make the script able to handle paths so that I can get the report also for subfolders, and for that I need to differentiate the desired folder from the rest of the path.
basename(1), dirname(1):
path=a/b/c
basename=$(basename "$path") # c
dirname=$(dirname "$path") # a/b
Prefix/suffix removal:
path=a/b/c
basename=${path##*/} # c
dirname=${path%/*} # a/b
Prefix/suffix removal is sufficient in some circumstances, and faster because it's native shell.
dirname/basename commands are slower (especially many paths in a loop etc) but handle more variable input or directory depth.
Eg. dirname "file" prints ., but suffix removal would print file. dirname /dir prints /, but suffix removal prints empty string; dirname also handles contiguous slashes (dirname a//b); basename a/b/ prints b, but prefix removal prints empty string.
If you know the structure is always 3 slashes (a/b/c), it may be safe to use prefix/suffix removal. But here I would use basename and dirname.
Also think about whether a better approach is to change the working directory with cd, so you can just refer to current directory with . (there's also $PWD and $OLDPWD).
Related
I want pick the specific format of file among the list of files in a directory. Please find the below example.
I have a below list of files (6 files).
Set-1
1) MAG_L_NT_AA_SUM_2017_01_20.dat
2) MAG_L_NT_AA_2017_01_20.dat
Set-2
1) MAG_L_NT_BB_SUM_2017_01_20.dat
2) MAG_L_NT_BB_2017_01_20.dat
Set-3
1) MAG_L_NT_CC_SUM_2017_01_20.dat
2) MAG_L_NT_CC_2017_01_20.dat
From the above three sets I need only 3 files.
1) MAG_L_NT_AA_2017_01_20.dat
2) MAG_L_NT_BB_2017_01_20.dat
3) MAG_L_NT_CC_2017_01_20.dat
Note: There can be multiple lines of commands because i have create the script for above req. Thanks
Probably easiest and least complex solution to your problem is combining find (a tool for searching for files in a directory hierarchy) and grep (tool for printing lines that match a pattern). You also can read those tools manuals by typing man find and man grep.
Before going straight to solution we need to understand, how we will approach your problem. To find pattern in a name of file we search we will use find command with option -name:
-name pattern
Base of file name (the path with the leading directories removed) matches shell pattern pattern. The metacharacters ('*', '?', and '[]')
match a '.' at the start of the base name (this is a change in
findutils-4.2.2; see section STANDARDS CONFORMANCE below). To ignore a
directory and the files under it, use -prune; see an example in the
description of -path. Braces are not recognised as being special,
despite the fact that some shells including Bash imbue braces with a
special meaning in shell patterns. The filename matching is performed
with the use of the fnmatch(3) library function. Don't forget to
enclose the pattern in quotes in order to protect it from expansion by
the shell.
For instance, if we want to search for a file containing string 'abc' in directory called 'words_directory', we will enter following:
$ find words_directory -name "*abc*"
And if we want to search all directories in directory:
$ find words_directory/* -name "*abc*"
So first, we will need to find all files, which begin with string "MAG_L_NT_" and end with ".dat", therefore to find all matching names in /your/specified/path/ which contains many subdirectories, which could contain files that match this pattern:
$ find /your/specified/path/* -name "MAG_L_NT_*.dat"
However this prints all found filenames, but we still get names containing "SUM" string, there comes in grep. To exclude names containing unwanted string we will use option -v:
-v, --invert-match
Invert the sense of matching, to select non-matching lines. (-v is
specified by POSIX .)
To use grep to filter out first commands output we will use pipe () |:
The standard shell syntax for pipelines is to list multiple commands,
separated by vertical bars ("pipes" in common Unix verbiage). For
example, to list files in the current directory (ls), retain only the
lines of ls output containing the string "key" (grep), and view the
result in a scrolling page (less), a user types the following into the
command line of a terminal:
ls -l | grep key | less
"ls -l" produces a process, the output (stdout) of which is piped to
the input (stdin) of the process for "grep key"; and likewise for the
process for "less". Each process takes input from the previous process
and produces output for the next process via standard streams. Each
"|" tells the shell to connect the standard output of the command on
the left to the standard input of the command on the right by an
inter-process communication mechanism called an (anonymous) pipe,
implemented in the operating system. Pipes are unidirectional; data
flows through the pipeline from left to right.
process1 | process2 | process3
After you got acquainted to mentioned commands and options which will be used to achieve your goal, you are ready for solution:
$ find /your/specified/path/* -name "MAG_L_NT_*.dat" | grep -v "SUM"
This command will produce output of all names which begin "MAG_L_NT_" and end with ".dat". grep -v will use first command output as input and remove all lines containing "SUM" string.
I am trying to iterate through every file in a specific directory (called sequences), and perform two functions on each file. I know that the functions (the 'blastp' and 'cat' lines) work, since I can run them on individual files. Ordinarily I would have a specific file name as the query, output, etc., but I'm trying to use a variable so the loop can work through many files.
(Disclaimer: I am new to coding.) I believe that I am running into serious problems with trying to use my file names within my functions. As it is, my code will execute, but it creates a bunch of extra unintended files. This is what I intend for my script to do:
Line 1: Iterate through every file in my "sequences" directory. (All of which end with ".fa", if that is helpful.)
Line 3: Recognize the filename as a variable. (I know, I know, I think I've done this horribly wrong.)
Line 4: Run the blastp function using the file name as the argument for the "query" flag, always use "database.faa" as the argument for the "db" flag, and output the result in a new file that is has the same name as the initial file, but with ".txt" at the end.
Line 5: Output parts of the output file from line 4 into a new file that has the same name as the initial file, but with "_top_hits.txt" at the end.
for sequence in ./sequences/{.,}*;
do
echo "$sequence";
blastp -query $sequence -db database.faa -out ${sequence}.txt -evalue 1e-10 -outfmt 7
cat ${sequence}.txt | awk '/hits found/{getline;print}' | grep -v "#">${sequence}_top_hits.txt
done
When I ran this code, it gave me six new files derived from each file in the directory (and they were all in the same directory - I'd prefer to have them all in their own folders. How can I do that?). They were all empty. Their suffixes were, ".txt", ".txt.txt", ".txt_top_hits.txt", "_top_hits.txt", "_top_hits.txt.txt", and "_top_hits.txt_top_hits.txt".
If I can provide any further information to clarify anything, please let me know.
If you're only interested in *.fa files I would limit your input to only those matching files like this:
for sequence in sequences/*.fa;
do
I can propose you the following improvements:
for fasta_file in ./sequences/*.fa # ";" is not necessary if you already have a new line for your "do"
do
# ${variable%something} is the part of $variable
# before the string "something"
# basename path/to/file is the name of the file
# without the full path
# $(some command) allows you to use the result of the command as a string
# Combining the above, we can form a string based on our fasta file
# This string can be useful to name stuff in a clean manner later
sequence_name=$(basename ${fasta_file%.fa})
echo ${sequence_name}
# Create a directory for the results for this sequence
# -p option avoids a failure in case the directory already exists
mkdir -p ${sequence_name}
# Define the name of the file for the results
# (including our previously created directory in its path)
blast_results=${sequence_name}/${sequence_name}_blast.txt
blastp -query ${fasta_file} -db database.faa \
-out ${blast_results} \
-evalue 1e-10 -outfmt 7
# Define a file name for the top hits
top_hits=${sequence_name}/${sequence_name}_top_hits.txt
# alternatively, using "%"
#top_hits=${blast_results%_blast.txt}_top_hits.txt
# No need to cat: awk can take a file as argument
awk '/hits found/{getline;print}' ${blast_results} \
| grep -v "#" > ${sequence_name}_top_hits.txt
done
I made more intermediate variables, with (hopefully) meaningful names.
I used \ to escape line ends and allow putting commands in several lines.
I hope this improves code readability.
I haven't tested. There may be typos.
You should be using *.fa if you only want files with a .fa ending. Additionally, if you want to redirect your output to new folders you need to create those directories somewhere using
mkdir 'folder_name'
then you need to redirect your -o outputs to those files, something like this
'command' -o /path/to/output/folder
To help you test this script out, you can run each line one by one to test them. You need to make sure each line works by itself before combining.
One last thing, be careful with your use of colons, it should look something like this:
for filename in *.fa; do 'command'; done
I'm a student and I'm very new to bash, so any help is greatly appreciated!
I'm trying to rename a batch of files that look like this: local_date_1415+6556_0001.txt and local_date_1415+6556_0002.txt.
Example file name: uuw_07052006_1415+6556_0001.txt
I need the "1415+6556" section of each filename to have a 2M in front of it, like "2M1415+6556". About half the files in the folder already have the 2M, so I can't just search for the string and replace.
Is there a way to rename the batch of files using "_" as a delimiter so I could replace all the third sections entirely with the correct string?
I have the rename command on my machine, I'm just not sure how to use it here.
Using your version of rename:
rename _ % *_????+*.txt # replace the first underscore with a percent
rename _ _2M *_????+*.txt # add 2M after the second underscore
rename % _ *_2M????+*.txt # return the first underscore back
Only works if your filenames don't contain %. If they do, pick a different character.
You can also write the loop yourself:
#! /bin/bash
for f in *_????+*.txt ; do
before=${f%[0-9][0-9][0-9][0-9]+*}
after=${f#*_[0-9][0-9][0-9][0-9][0-9][0-9][0-9][0-9]}
mv "$f" "$before"2M"$after"
done
I need to unzip a bunch of student assignment (jar) files so that I can use a script to submit the contents to the Moss (Stanford) plagiarism detection server. I did the same thing in Java which was trivial but I'm trying to re-implement to as a bash script.
I am trying to do the following:
Get a list of student names (each student has a directory).
In each student directory, sub-directories exist numbered from 1 to the
latest submission. I need to get the directory with the highest
number.
Inside of each of those submission directories contains a
jar file that I need. I copy each jar into a temp directory with the
same name as the student and unzip it.
I need that temp directory listing formatted as a string in the form
/tempDir/studentName1/.languageExt /tempDir/studentName2/.languageExt
The student directory has the basic structure:
Student_Root_Directory:
Student1
Student2
Student1
Sub-Directories: 1 2 3 4 5
1: student1.jar
2: student1.jar
...
Student2
Sub-Directories: 1 2 3
1. student2.jar
...
To do the first 3 steps above I did:
#!/bin/bash
# Extract all jar files into a temp directory called /home/moss/tempJarFiles/studentName
# $1 is the command line argument that contains the path to the institution submission dir.
# $2 is the language extension: .c, .cpp, .java, .py
students=`ls $1`
student_dir=$1
languageExt=$2
mossDir="/home/moss"
tempDir="/home/moss/tempJarStorage"
for student in $students
do
latestSubmissionDir=`ls -t $student_dir/$student | head -1`
for jarDir in $latestSubmissionDir
do
mkdir $tempDir/$student
cp $student_dir/$student/$jarDir/*.jar $tempDir/$student
unzip -d $tempDir/$student/ -o -j $tempDir/$student/$student.jar *.$languageExt
rm $tempDir/$student/$student.jar
done
done
...which results in a number of student directories being created in a temp directory that contains only the unzipped contents for the student submissions.
I need the ls output of the new temp directories formatted as a string that contains:
/tempDir/studentName1/\*.languageExt /tempDir/studentName2/\*.languageExt
I have tried variations on
find "$tempDir" -iname "*.$languageExt" -printf "%p/*.$languageExt"
using iname and not - but I either have output that contains extra directory information such as $tempDir/*.languageExt (when I just need the subdirectories $tempDir/$studentName/*.languageExt) or I have output where the path for every source file is also listed such as:
$tempDir/$studentName/studentNameA.java
$tempDir/$studentName/studentNameB.java
when I only need
$tempDir/$studentName/*.java
I think this should be really easy and I'm just over thinking it. Any hints for improving the script also appreciated.
Here's a revised version of the script hat may work:
#/bin/bash
# Extract all jar files into a temp directory called /home/moss/tempJarFiles/studentName
# $1 is the command line argument that contains the path to the institution submission dir.
# $2 is the language extension: c, cpp, java, py
students_dir=$1
languageExt=$2
studentPathsT=( "$students_dir"/*/ )
mossDir='/home/moss'
tempDir='/home/moss/tempJarStorage'
for studentPathT in "${studentPathsT[#]}"; do
student=$(basename "$studentPathT")
mkdir "$tempDir/$student"
submissionDirsT=( "$studentPathT"*/ )
latestSubmissionDirT=${submissionDirsT[${#submissionDirsT[#]-1]}
cp "$latestSubmissionDirT"*.jar "$tempDir/$student/"
unzip -d "$tempDir/$student/" -o -j "$tempDir/$student/*.jar" "*.$languageExt"
rm "$tempDir/$student"/*.jar
done
# Note that at this point `"$tempDir"/*/*.$languageExt` would expand
# to all extracted submission files, across all students.
# Finally, output each student's extracted files as an unexpanded glob à la
# /{tempDir}/{studentName1}/*.{languageExt}
for pT in "$tempDir"/*/; do
echo "$pT*.$languageExt"
# Note: If there is a chance that your filenames contain
# embedded newlines (rare in practice) using `echo` won't work properly
# as #Charles Duffy points out.
# If that is a concern, use
# printf '%s\0' "$pT*.$languageExt"
# and process the output with a utility that can process NUL characters
# as separators, such as `xargs -0`.
done
It avoids using ls and only uses pathname expansion and array variables so as to properly deal with paths that contain embedded spaces and other shell metacharacters.
suffix ...T in variable names indicates that a particular path or array of paths is *T*erminated, i.e, that it ends in a /.
The assumption is that the numbered subdirectories do not go beyond 9, as the implicit lexical sorting of pathname expansion is relied upon; if the numbers go higher, explicit numerical sorting must be applied.
Note that the globs (pathname patterns) passed to unzip are intentionally double-quoted, as they should be interpreted by unzip, not the shell.
Note that, based on your original code, I've assumed that $languageExt does NOT start with . (e.g., cpp rather than .cpp), despite what your comment says.
I am writing a bash script that will output a .tgz file to a specific directory, /tmp/ by default
I would like to provide an option to override this directory and I have chosen to do so using arguments provided at the command line
while getopts d: option
do
case "${option}" in
d) dir=${OPTARG};;
esac
done
As written, this works but I've run into a snag depending on user input
The name of my .tgz file is also a variable and my code that brings this all together is
output="$dir""$name"
The problem that I run into is if the user runs
./script -d /home/user
My resulting path and filename end up as
/home/userfilename.tgz
I need to either enforce a requirement for a trailing / or insert one if the user did not.
While it works, if I change my output variable to
output="$dir"/"$name"
If the user does provide a trailing / I end up with something like this and I am trying to keep my output aesthetic.
/home/user//filename.tgz
Any input would be greatly appreciated.
Add the line
output="${output//\/\///}"
after joining dir and name.
It looks complicated, but what it does is it replaces two slashes with one.
You may find more info in here.