"awk with if / substr " didn't work as expected

Question

I want to filter the records where disk usage is greater than 70% using df and awk/if/substr. The problem is it looks like only the first character is considered when comparing:

commands:

# df -Pah | awk '{if(substr($5,1,length($5)-1)>70) print $5}'
Use%
9%
8%

But like this it works:

commands:

# df -Pah | awk '{if(substr($5,1,length($5)-1)-70>0) print $5}'
100%
100%
100%
# df -Pah | awk '{if(substr($5,1,length($5)-1)-30>0) print $5}'
54%
35%
100%
100%
100%

Does anyone know why?

PS: df -Pah output, in fact it's quit the same as df -h:

Filesystem  Size  Used Avail Use% Mounted on
/dev/sda2   2.0G 1020M  894M  54% /    
udev        4.0G  420K  4.0G   1% /dev
tmpfs       4.0G   16M  3.9G   1% /dev/shm

Answer 1

Awk comparisons are as I recall either numeric or string comparisons, depending on the type of the left operand, and that would explain what you're seeing: substr gives a string result, and strings beginning 'U', '9', and '8' are all greater than any string beginning '7'.

Subtraction produces a numeric result.