Fix the file handles issue discussed in #5066

[kgururaj]

Some speedup by eliminating a ls operation - however, having 10K small arrays will still be slower than 100 large arrays.

[codecov-io]

Codecov Report

Merging #5105 into master will increase coverage by 1.151%.
The diff coverage is n/a.

@@               Coverage Diff               @@
##              master     #5105       +/-   ##
===============================================
+ Coverage     85.341%   86.492%   +1.151%     
- Complexity     28712     29204      +492     
===============================================
  Files           1814      1814               
  Lines         135364    135364               
  Branches       15042     15042               
===============================================
+ Hits          115521    117079     +1558     
+ Misses         14309     12825     -1484     
+ Partials        5534      5460       -74

Impacted Files	Coverage Δ	Complexity Δ
...utils/smithwaterman/SmithWatermanIntelAligner.java	50% <0%> (-30%)	1% <0%> (-2%)
...ithwaterman/SmithWatermanIntelAlignerUnitTest.java	60% <0%> (ø)	2% <0%> (ø)	⬇️
...nstitute/hellbender/utils/read/AlignmentUtils.java	75.605% <0%> (+0.202%)	149% <0%> (+1%)	⬆️
...ools/walkers/haplotypecaller/graphs/BaseGraph.java	82.239% <0%> (+0.386%)	94% <0%> (+1%)	⬆️
...roadinstitute/hellbender/utils/read/ReadUtils.java	80.653% <0%> (+0.466%)	209% <0%> (ø)	⬇️
...kers/variantutils/PosteriorProbabilitiesUtils.java	81.41% <0%> (+0.641%)	50% <0%> (ø)	⬇️
...nder/utils/runtime/StreamingProcessController.java	72.249% <0%> (+0.957%)	36% <0%> (+1%)	⬆️
...annotator/allelespecific/AS_RMSMappingQuality.java	88.35% <0%> (+0.971%)	34% <0%> (ø)	⬇️
.../utils/smithwaterman/SmithWatermanJavaAligner.java	94.118% <0%> (+1.176%)	46% <0%> (ø)	⬇️
...ols/walkers/haplotypecaller/AssemblyResultSet.java	75.449% <0%> (+1.198%)	44% <0%> (+1%)	⬆️
... and 49 more

[droazen]

👍