Github user paul-rogers commented on a diff in the pull request:
https://github.com/apache/drill/pull/1060#discussion_r158196570
--- Diff: exec/vector/src/main/codegen/templates/NullableValueVectors.java
---
@@ -68,96 +85,441 @@
private final UInt1Vector bits = new UInt1Vector(bitsField, allocator);
private final ${valuesName} values = new ${minor.class}Vector(field,
allocator);
+ private final Mutator mutator = new MutatorImpl();
+ private final Accessor accessor = new AccessorImpl();
+
+ <#if type.major == "VarLen" && minor.class == "VarChar">
+ private final Mutator dupMutator = new DupValsOnlyMutator();
+ /** Accessor instance for duplicate values vector */
+ private final Accessor dupAccessor = new DupValsOnlyAccessor();
+ /** Optimization for cases where all values are identical */
+ private boolean duplicateValuesOnly;
+ /** logical number of values */
+ private int logicalNumValues;
+ /** logical value capacity */
+ private int logicalValueCapacity;
--- End diff --
Explain this? If I have `"foo.csv"` as my value (for `filename`, say) what
does it mean to have a "capacity"? Can't I have 64K "copies" of my one value?
---
