Category: Software

Let’s say that you have a file and you would like to convert each line in the file to an element in an array.

The key to this is knowing about and how to manipulate the IFS (Internal Field Separator).  The default IFS is whitespace (a space, tab, or newline) and if you create an array passing it a whitespace delimited list of strings, each token will be set to an element in the array.

ARRAY=(a b d c)

Will result in an array with a single letter in each element.

To do the same thing with the contents of a file, whereby each element is on a separate line, the first thing to be done is to set the IFS that is just new-lines (carriage returns).  Then set, as the input for the array, the contents of the file.

# Save our existing IFS
OIFS="$IFS"

# Set our IFS to a new-line/carriage return
IFS=$'\r\n'

# Create the array with the contents of a file
TEST_ARRAY=($(cat some_file.txt))

# Reset our IFS
IFS="$OIFS"

for i in "${TEST_ARRAY[@]}"
do
   echo $i
done

About a year ago I posted an article about how to split into an array of values based on a given delimiter in bash.

The following is how to take that same string and split it on the first occurrence of the same user defined delimiter.

Both use the ‘read’ command, but in a slightly different way.

Instead of passing read the -a [aname] parameter which tells it that “The words are assigned to sequential indices of the array variable aname, starting at 0.”, we pass is -r which indicates that “Backslash does not act as an escape character.  The backslash is considered to be part of the line.”.  This will make sure to include any backslash that is in the string in your output.

Then, we provide two variables into which we will store the split string.

#!/bin/bash

SOURCE_STRING='foo|blah|moo'

# Save the initial Interal Field Separator
OIFS="$IFS"

# Set the IFS to a custom delimiter
IFS='|'

read -r KEY VALUE <<< "${SOURCE_STRING}"
echo "KEY = $KEY, VALUE = $VALUE"

# Reset original IFS
IFS="$OIFS"

Often times you will want to write a BASH script where you don’t want to have to keep track of all of the positional command line arguments and/or you might want to configure it with a set of environmental variables while having a default value for each in the script.

Following is the syntax for declaring them in the shell script, and then an example on how to invoke it.

#!/bin/bash

: ${ARG1:="somedefault_arg1"}
: ${ARG2:="10"}

echo "ARG1 = $ARG1"
echo "ARG2 = $ARG2"

$ ./default-bash-vars.sh
ARG1 = somedefault_arg1
ARG2 = 10
$ ARG1="someOtherArg1" ARG2="20" ./default-bash-vars.sh
ARG1 = someOtherArg1
ARG2 = 20

In the example script, we have two variables, ARG1 and ARG2.  When running the script without providing any additional configuration the default values will be used.  When invoking it and defining the variables on the command line prior to executing the script those values will be used instead.

This prevents the situation where you potentially have many command-line arguments and then have to jugle the positional $1, $2, …. vars in the script.