How to find number of unique 'applications' in Debian jessie

Question

This is a follow-up query on allpackages.txt.gz has two different file-sizes?

$ wc -l allpackages.txt 
57725 allpackages.txt

I looked at the output and it says 57725 -6 the header files which is the grand total of 57719 packages but as shared in the comments on http://linux.koolsolutions.com/2008/12/19/counting-number-of-packages-under-debian-linux/#comment-10014 the following command should work but doesn't.

$ tail -n +7 allpackages.txt | cut -d’ ‘ -f1 | cut -d’-‘ -f1 | uniq | wc -l
cut: cut: the delimiter must be a single character
Try 'cut --help' for more information.
the delimiter must be a single character
Try 'cut --help' for more information.
0

Can somebody help in couple of ways :-

a. Decoding what the command itself tries to do ? I know that tail is used to give the last 10 odd words and cut is when you are trying to remove some sections of the file.

If needed would share the allpackages.txt somewhere so people can figure out the best way to know the unique number of packages therein.

https://www.dropbox.com/sh/9p03q138xx9nxvb/AAAukls-UnxRZwoCPwx4HwLAa?dl=0

Answer 1

You showed the command: tail -n +7 allpackages.txt | cut -d’ ‘ -f1 | cut -d’-‘ -f1 | uniq | wc -l. This contains non-ASCII quotes, which certainly isn't what was intended.

Try this instead:

tail -n +7 allpackages.txt | cut -d' ' -f1 | cut -d- -f1 | uniq | wc -l

Here's what the members of the pipeline do:

• tail -n +7 prints a file starting at line 7;

• cut -d' ' -f1 keeps the first field, using a single space as a separator;

• cut -d- -f1 keeps the first field, using a minus sign (-) as a separator;

• uniq compresses runs of identical lines into a single line;

• wc -l prints the line count.

So, the entire pipeline ignores the first six lines of your file, and then attempts to count the unique values at the selected field. (This assumes that the input is sorted so that the values to be squashed together come in sequence, if this isn't true, a call to sort, perhaps sort -u, could be used).