Open LettuceBacon opened 1 year ago
Hello. I changed line 86 in def_global.h to "typedef long long Timestamp;" and tested a small dataset containing only two intervals: 1640966444000 1640966934000 0 231232144347
and a query file containing 0 4
Num of intervals : 2 Domain size : 1640966934000 Avg interval extent [%] : 7.045621
Optimizations : no Num of bits : 3 Num of partitions : 15 Num of Originals : 2 Num of replicas : 0 Num of empty partitions : 13 Avg partition size : 1.000000 Read VM [Bytes] : 0 Read RSS [Bytes] : 0 Indexing time [secs] : 0.000045
Predicate type : GOVERLAPS Strategy : bottom-up Num of runs per query : 1 Num of queries : 1 Avg query extent [%] : 0.000000 Total result [XOR] : 1 Total querying time [secs]: 0.000000 Avg querying time [secs] : 0.000000
Throughput [queries/sec] : inf
Everything seems to be working; the throughput is of course "inf" because the total query time is 0; small bug there, not checking before the division.
Thanks for reply!
After changing the typedef of "Timestamp", I got proper result on the test case you mentioned. But I got "Segement fault" when ran other test cases without integer endpoints. Here's are the details.
Dataset only containing intervals withlong long
endpoints:
1640997298000 1640999660000
1640997297000 1640999680000
Query file only containing intervals withlong long
endpoints:
1640997288000 1640999760000
1640997297000 1640999680000
Run cmd:
./query_hint_m.exec -m 3 -t -q gOVERLAPS test.dat test.qry
Result:
Segmentation fault
Dataset containing intervals with long long
and int
endpoints:
4 12345
1640997297000 1640999680000
Query file only containing intervals withlong long
endpoints:
1640997288000 1640999760000
1640997297000 1640999680000
Run cmd:
./query_hint_m.exec -m 3 -t -q gOVERLAPS test.dat test.qry
Result:
HINT^m
======
Num of intervals : 2
Domain size : 1640999679996
Avg interval extent [%] : 0.000073
Optimizations : no
Num of bits : 3
Num of partitions : 15
Num of Originals : 2
Num of replicas : 0
Num of empty partitions : 13
Avg partition size : 1.000000
Read VM [Bytes] : 0
Read RSS [Bytes] : 0
Indexing time [secs] : 0.000012
Predicate type : GOVERLAPS
Strategy : top-down
Num of runs per query : 1
Num of queries : 2
Avg query extent [%] : 0.000148
Total result [XOR] : 2
Total querying time [secs]: 0.000000
Avg querying time [secs] : 0.000000
Throughput [queries/sec] : 6060606.060606
Dataset only containing intervals withlong long
endpoints:
1640997298000 1640999660000
1640997297000 1640999680000
Query file containing intervals withlong long
and int
endpoints:
4 12345
1640997297000 1640999680000
Run cmd:
./query_hint_m.exec -m 3 -t -q gOVERLAPS test.dat test.qry
Result:
Segmentation fault
Dataset containing intervals in different order from Test case 2:
1640997298000 1640999660000
4 12345
Query file containing intervals withlong long
and int
endpoints:
1640997298000 1640999660000
1640997297000 1640999680000
Run cmd:
./query_hint_m.exec -m 3 -t -q gOVERLAPS test.dat test.qry
Result:
HINT^m
======
Num of intervals : 2
Domain size : 1640999659996
Avg interval extent [%] : 0.000072
Optimizations : no
Num of bits : 3
Num of partitions : 15
Num of Originals : 2
Num of replicas : 0
Num of empty partitions : 13
Avg partition size : 1.000000
Read VM [Bytes] : 0
Read RSS [Bytes] : 0
Indexing time [secs] : 0.000010
Predicate type : GOVERLAPS
Strategy : top-down
Num of runs per query : 1
Num of queries : 2
Avg query extent [%] : 0.000145
Total result [XOR] : 0
Total querying time [secs]: 0.000000
Avg querying time [secs] : 0.000000
Throughput [queries/sec] : 9049773.755656
Hello. I am trying to use this index to process intervals with long type end points. A record likes
1640966444000 1640966934000
. Execute cmd in README.md leads to a weird result. "Avg interval extent" and "Avg partition size" are "-nan". "Indexing time" is so short (0.000060) that it looks like it wasn't executed. Then I change line 86 of def_global.h into "typedef long Timestamp;" but get an error like below.How should I edit the code to support long type values?