The method added in the PR is needed for S3 DatasetScans and also if we want featureCollection indices in S3. The idea is that if we have a directory and want to get the file path of a file in that directory, we should let MFile handle that to avoid repeatedly parsing path strings (For S3 the separator could be "/" or "?" if its a bucket, and it also may have a fragment in the uri like cdms3:mybucket/dir#delimiter=/ so appending /filename.nc or whatever doesn't work.)
Description of Changes
The method added in the PR is needed for S3 DatasetScans and also if we want featureCollection indices in S3. The idea is that if we have a directory and want to get the file path of a file in that directory, we should let
MFilehandle that to avoid repeatedly parsing path strings (For S3 the separator could be "/" or "?" if its a bucket, and it also may have a fragment in the uri likecdms3:mybucket/dir#delimiter=/so appending/filename.ncor whatever doesn't work.)