If you were to use the whole string as the key to the hash, what would it buy you?
It would make for problems in building the data structure because for the compound lines, you wouldn't have the entire key when you read the first line. You'd either have to employ some readahead, or delete and re-store compound lines under new keys each time you found an extension line.
It would be possible to sort the data prior to spliting it, but if the fields are complex (like dates) then it's much easier to do the sort after the split.
Without making any attempt to be efficient, sorting by field n (a more normal term for your d(n) nomenclature), can be very simple. This sorts the data by the (additional) last character of the 4th field:
#! perl -slw use strict; my @data; while( <DATA> ) { if( /^\s/ ) { push @{ $data[ -1 ] }, split; } else { push @data, [ split ]; } } print "@$_" for sort{ substr( $a->[ 3 ], -1 ) cmp substr( $b->[ 3 ], -1 ) }@data; __DATA__ a1q b1w c1e d1r e1t f1y a2u b2i c2o d2p e2a f2s a3d b3f c3g d3h e3j f3k p3 q3 r3 s3 t3 u3 a4l b4z c4x d4c e4v f4b a5n b5m c5q d5w e5e f5r p5 q5 r5 s5 t5 u5 a6t b6y c6u d6i e6o f6p
Sorting by a date field is slightly more complex, but not much. I'd give an example, but as you've given dummy data, I'd have to make up the dates and I've no idea what format your data is in.
If you posted an example of your real data and explained what you are actually trying to achieve, rather than all this abstract stuff, you'd doubtless get much better answers.
In reply to Re^3: Selecting the right data structure
by BrowserUk
in thread Selecting the right data structure
by thezip
| For: | Use: | ||
| & | & | ||
| < | < | ||
| > | > | ||
| [ | [ | ||
| ] | ] |