Question

I have written a small bash* script to process files in a directory. The files have a specific naming convention, which determines how the file is to be processed.

The filename starts with a geographic origin, and then contains a keyword in the filename that determines how it should be processed.

Here is the stripped down version of my script:

DATA_DIRECTORY=./data

array=( 'uk' 'usa' 'fra' 'ukr' )

for i in "${array[@]}"
do
    echo "#################################################"
    echo "Loop index : $i"
    echo "#################################################"

    # process baboon data for the current country
    for filename in $DATA_DIRECTORY/$i_*_baboon*.csv; do
       echo "Processing file: $filename in loop: $i"
    done

    # process whale data for the current country
    for filename in $DATA_DIRECTORY/$i_*_whale*.csv; do
       echo "Processing file: $filename in loop: $i"
    done

    # process lion data for the current country
    for filename in $DATA_DIRECTORY/$i_*_lion*.csv; do
       echo "Processing file: $filename in loop: $i"
    done

done

I am expecting the script above to match the files in the following way: A file named us_2007_deaths_caused_by_baboon.csv will be processed once – when the loop counter variable is ‘us’ and during the loop for filename in $DATA_DIRECTORY/$i_*_baboon*.csv;. However, the matching is not working as I expected, and each (nested for loop) is looping over all of the files in the directory – which is NOT what I want.

How may I pattern match the files based on the filenames?

*Note: I am using bash version 4.1.5 on Ubuntu 10.0.4

Answer 1

Write ${i}_ instead of $i_ because $i_ is “” (bash thinks that i_ is a variable here).

For example:

for filename in $DATA_DIRECTORY/${i}_*_baboon*.csv; do
   echo "Processing file: $filename in loop: $i"
done