Github user hvanhovell commented on a diff in the pull request: https://github.com/apache/spark/pull/14083#discussion_r69984539 --- Diff: sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/plans/logical/LogicalPlan.scala --- @@ -165,111 +169,99 @@ abstract class LogicalPlan extends QueryPlan[LogicalPlan] with Logging { def resolveQuoted( name: String, resolver: Resolver): Option[NamedExpression] = { - resolve(UnresolvedAttribute.parseAttributeName(name), output, resolver) + outputAttributeResolver.resolve(UnresolvedAttribute.parseAttributeName(name), resolver) } /** - * Resolve the given `name` string against the given attribute, returning either 0 or 1 match. - * - * This assumes `name` has multiple parts, where the 1st part is a qualifier - * (i.e. table name, alias, or subquery alias). - * See the comment above `candidates` variable in resolve() for semantics the returned data. + * Refreshes (or invalidates) any metadata/data cached in the plan recursively. */ - private def resolveAsTableColumn( - nameParts: Seq[String], - resolver: Resolver, - attribute: Attribute): Option[(Attribute, List[String])] = { - assert(nameParts.length > 1) - if (attribute.qualifier.exists(resolver(_, nameParts.head))) { - // At least one qualifier matches. See if remaining parts match. - val remainingParts = nameParts.tail - resolveAsColumn(remainingParts, resolver, attribute) - } else { - None - } + def refresh(): Unit = children.foreach(_.refresh()) +} + +/** + * Helper class for (LogicalPlan) attribute resolution. This class indexes attributes by their + * case-in-sensitive name, and checks potential candidates using the given Resolver. Both qualified --- End diff -- The `resolve` methods takes a `Resolver` as its parameter. This allows us to use either case sensitive or insensitive attribute resolution depending on the `Resolver` passed. The names of both classes are confusing and I might rename the `AttributeResolver` class to `AttributeIndex` or something like that... The `AttributeResolver` creates two indexes based on the lower case (qualified) attribute name; we do an initial lookup based on the lower case name, and then use the `Resolver` for the actual attribute selection. This allows us to do fast(er) and correct lookups.
--- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. --- --------------------------------------------------------------------- To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org