Perhaps this is explained deep in the shop manual somewhere, but I
decided to be lazy and try asking here first.
I'd like to have a *Boolean* array indexed by an array-of-bytes, or
perhaps a string. This can be accomplished with, e.g.,
JHSI(PValue, PJHSArray, data, N);
if (*PValue !=0 ) {
Code to run when the N bytes from
data were previously seen.
} else {
*PValue=1;
Code to run when the data is new.
}
This allocates an 8-byte integer which is set to 1 in addition to
recording the data string itself. If my data strings are relatively
short, perhaps 10 or 20 bytes, would it be faster and/or smaller to
"roll my own" version which would create a tree with one or more JudyL
levels at the top and Judy1 trees at the bottom? Or, would this not be
worth it? In the application I'm looking at right now, the strings have
a fixed length, which means I can determine the geometry in advance.
Also, now that 128 bit integers are fairly common, would it be possible,
easy/hard, advisable/or-not, to create a 128 bit version of Judy? Then
Judy1 would be able to handle a 16-byte string.
Thanks in advance,
-Jeff Norden, [email protected]
Dept of Math, Tenn Tech Univ, Cookeville TN 38505 (931)372-3441
PS: hope everyone reading this is safe and healthy.
_______________________________________________
Judy-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/judy-devel