Extract filename from another filename

Question

I receive notification files for different files. I want to extract the filename from these notification files which I would use in another process. The first 4 strings of the notification file will always be ignored. So the filename starts at the 5 position, can start with any character and ends at the .csv. The filename could be any length.

An example:

NotificationFile = 'NFT_PPF_STANDARD_1720156B_NBIRNans_DFT_TILT_EOF_FFP_20160619-10.csv_16062000352520'
NotificationFile = 'NFT_PPF_IMPULSE_172555A_Acc_Account_FFP_20160619-10.csv_16062000352520'
NotificationFile = 'NFT_PPF_IMPULSE_172555A_CRI_Position_FFP_20160619-10.csv_16062000352520'

The filenames respectively:

Filename = 'NBIRNans_DFT_TILT_EOF_FFP_20160619-10.csv'
Filename = 'Acc_Account_FFP_20160619-10.csv'
Filename = 'CRI_Position_FFP_20160619-10.csv'

I use the following command to extract the filename from the notification filename but it only works if the filename contains only 3 '_'s:

Extract = echo $NotificationFile | cut -d "_" -f5-8
# $NotificationFile is the NotificationFile above
# Extract should be equal to Filename 

What is a general command I can use for any filename length?

Answer 1

For what you want, you could use sed. This is how I would do it.

sed -ne 's/.*\(NB.*\)\_.*$/\FILENAME = \1 /p' 

Answer 2

If all the filenames end with .csv, and don't contain another .csv anywhere, use eg. sed to remove everything after that:

Extract=$(echo $NotificationFile | sed 's/\.csv.*/.csv/')

Combining that with what you had before:

Extract=$(echo $NotificationFile | sed 's/\.csv_.*/.csv/' | cut -d "_" -f5-)

The range N- means "to end of line", ie. not stopping at the 8th if there weren't exactly 3 _s in the filename.

Answer 3

Use the string manipulation constructs of parameter expansion.

filename="${NotificationFile#????}"
filename="${filename%.csv*}.csv"

The first line sets filename to the value of NotificationFile with the first four characters stripped off. The second line removes everything from filename starting with .csv, and re-adds .csv to compensate.