BWASW more reads in the output SAM file than in the input file

[nanto]

I also heard that with extremely long reads bwa-sw can give not very reliable output. Is there any sense in browsing reads up to a given length, or it will change nothing, at will only consume my time.

[gaffa]

BWA-SW performs local alignment and can output several alignments per input read/sequence. So a given read or different parts of a read can be aligned to different places, which are all included in the output file.

[nanto]

When running BWA I get normal total number of reads (the same as in the input file).

But when running BWASW alignment I get this:

(that's from samtools flagstat)

56125 + 0 in total (QC-passed reads + QC-failed reads)
0 + 0 duplicates
30118 + 0 mapped (53.66%:-nan%)
0 + 0 paired in sequencing
0 + 0 read1
0 + 0 read2
0 + 0 properly paired (-nan%:-nan%)
0 + 0 with itself and mate mapped
0 + 0 singletons (-nan%:-nan%)
0 + 0 with mate mapped to a different chr
0 + 0 with mate mapped to a different chr (mapQ>=5)

My input number of reads loaded to BWA was 45997. How come I got over 80k?