gzip files with cat and from pipe gives different results

Question

Why does this give such output (both commands are supposed to do the same thing) and how can one make them give identical output?

diff <(cat some_file | gzip -c - | base64) <(gzip -c some_file | base64)
1,2c1,2
< H4sIACSOZFUAA2XNsRHAMAgDwDqZRkIQ8P6L+c5xnIL2m2c5E6BdIQA5cHPTaGTqlI3ki2jSoWrk
< e1Tw0PNSMT4KdPKfJgNiJT++AAAA
---
> H4sICGcqSlUAA2Z0X2FkLnNob3J0AGXNsRHAMAgDwDqZRkIQ8P6L+c5xnIL2m2c5E6BdIQA5cHPT
> aGTqlI3ki2jSoWrke1Tw0PNSMT4KdPKfJgNiJT++AAAA

The contents of the file are:

184170012   53000790
184170019   53000790
184170023   53000790
184170027   53000790
184170034   53001233
184170038   53001233
184170042   53000351
184170046   53000815
184170050   53000815
184170054   53000815

There is a tab character between two columns and new line at the end of each line.

Answer 1

gzip is encoding the filename of the input file into its output. Even with -c option it does this. You can see this with gzip -c some_file | strings|head -1. however, when reading from stdin, gzip does not do that, since it doesn't know the filename. You can tell gzip to omit from output the filename and time-stamp with -n.