abhishekrb19 commented on code in PR #14319:
URL: https://github.com/apache/druid/pull/14319#discussion_r1200950594
##########
sql/src/main/java/org/apache/druid/sql/calcite/schema/SegmentMetadataCache.java:
##########
@@ -995,4 +986,117 @@ void doInLock(Runnable runnable)
runnable.run();
}
}
+
+
+ /**
+ * ColumnTypeMergePolicy defines the rules of which type to use when faced
with the possibility of different types
+ * for the same column from segment to segment. It is used to help compute a
{@link RowSignature} for a table in
+ * Druid based on the segment metadata of all segments, merging the types of
each column encountered to end up with
+ * a single type to represent it globally.
+ */
+ @JsonTypeInfo(use = JsonTypeInfo.Id.NAME, property = "type", defaultImpl =
FirstTypeMergePolicy.class)
+ @JsonSubTypes(value = {
+ @JsonSubTypes.Type(name = FirstTypeMergePolicy.NAME, value =
FirstTypeMergePolicy.class),
+ @JsonSubTypes.Type(name = LeastRestrictiveTypeMergePolicy.NAME, value =
LeastRestrictiveTypeMergePolicy.class)
+ })
+ @FunctionalInterface
+ public interface ColumnTypeMergePolicy
+ {
+ ColumnType merge(ColumnType existingType, ColumnType newType);
+ }
+
+ /**
+ * Classic logic, we use the first type we encounter. This policy is
effectively 'newest first' because we iterated
+ * segments starting from the most recent time chunk, so this typically
results in the most recently used type being
+ * chosen, at least for systems that are continuously updated with 'current'
data.
+ *
+ * Since {@link ColumnTypeMergePolicy} are used to compute the SQL schema,
at least in systems using SQL schemas which
+ * are poartially or fully computed by this cache, this merge policy can
result in query time errors if incompatible
+ * types are mixed if the chosen type is more restrictive than the types of
some segments. If data is likely to vary
+ * in type across segments, consider using {@link
LeastRestrictiveTypeMergePolicy} instead.
+ */
+ public static class FirstTypeMergePolicy implements ColumnTypeMergePolicy
Review Comment:
That makes sense. Thanks for clarifying.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
