You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Copying the complete collection when calling .par on IndexedSeqs is wasteful. There are use cases where the IndexedSeq is backed by an external data store which may be (much) bigger than what currently fits into memory. Is there any rationale for that copying in the case of IndexedSeqs? Or is there just no alternative implementation yet?
The text was updated successfully, but these errors were encountered:
@axel22 said:
Good question. Well, IndexedSeq still allows the apply operation to be O(log n). But I guess we could have a default implementation for the indexed sequences to avoid the copying, and have them borrow combiners from a concrete collection such as a parallel array.
Copying the complete collection when calling
.par
onIndexedSeqs
is wasteful. There are use cases where theIndexedSeq
is backed by an external data store which may be (much) bigger than what currently fits into memory. Is there any rationale for that copying in the case ofIndexedSeqs
? Or is there just no alternative implementation yet?The text was updated successfully, but these errors were encountered: