Github user marmbrus commented on a diff in the pull request:
https://github.com/apache/spark/pull/2475#discussion_r19651082
--- Diff:
sql/core/src/main/scala/org/apache/spark/sql/sources/package.scala ---
@@ -0,0 +1,95 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF) under one or more
+ * contributor license agreements. See the NOTICE file distributed with
+ * this work for additional information regarding copyright ownership.
+ * The ASF licenses this file to You under the Apache License, Version 2.0
+ * (the "License"); you may not use this file except in compliance with
+ * the License. You may obtain a copy of the License at
+ *
+ * http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+package org.apache.spark.sql
+
+import org.apache.spark.annotation.DeveloperApi
+import org.apache.spark.rdd.RDD
+import org.apache.spark.sql.catalyst.expressions.{Expression, Attribute}
+
+/**
+ * A set of APIs for adding data sources to Spark SQL.
+ */
+package object sources {
+
+ /**
+ * Implemented by objects that produce relations for a specific kind of
data source. When
+ * Spark SQL is given a DDL operation with a USING clause specified,
this interface is used to
+ * pass in the parameters specified by a user.
+ *
+ * Users may specify the fully qualified class name of a given data
source. When that class is
+ * not found Spark SQL will append the class name `DefaultSource` to the
path, allowing for
+ * less verbose invocation. For example, 'org.apache.spark.sql.json'
would resolve to the
+ * data source 'org.apache.spark.sql.json.DefaultSource'
+ */
+ trait RelationProvider {
+ /** Returns a new base relation with the given parameters. */
+ def createRelation(sqlContext: SQLContext, parameters: Map[String,
String]): BaseRelation
+ }
+
+ /**
+ * Represents a collection of tuples with a known schema. Classes that
extend BaseRelation must
+ * be able to produce the schema of their data in the form of a
[[StructType]] In order to be
+ * executed, a BaseRelation must also mix in at least one of the Scan
traits.
+ *
+ * BaseRelations must also define a equality function that only returns
true when the two
+ * instances will return the same data. This equality function is used
when determining when
+ * it is safe to substitute cached results for a given relation.
+ */
+ @DeveloperApi
+ abstract class BaseRelation {
+ def sqlContext: SQLContext
+ def schema: StructType
+ }
+
+ /**
+ * Mixed into a BaseRelation that can produce all of its tuples as an
RDD of Row objects.
+ */
+ @DeveloperApi
+ trait TableScan {
+ self: BaseRelation =>
+
+ def buildScan(): RDD[Row]
+ }
+
+ /**
+ * Mixed into a BaseRelation that can eliminate unneeded columns before
producing an RDD
+ * containing all of its tuples as Row objects.
+ */
+ @DeveloperApi
+ trait PrunedScan {
+ self: BaseRelation =>
+
+ def buildScan(requiredColumns: Seq[Attribute]): RDD[Row]
--- End diff --
To be clear they will always be resolved here (or the analyzer would have
thrown an exception). I just usually don't use AttributeReference in type
signatures so as to hide their implementation details.
You raise some good points though, and are making me reconsider an earlier
design where we don't expose `Attribute` or `Expression`. Instead we have an
interface like this:
Attributes are `String`. This can just be `attributeName`. We can possibly
also support `attributeName.fieldName` (or maybe even `attributeName[0]`?).
The planner will default to just pushing down full attribute projections, but
the developer can override methods to turn on the more advanced projections.
For filters we define a simple set initially that we can add to as we go:
```scala
case class EqualTo(attribute: String, value: Any)
```
Pros: Easier to make binary compatibility promises. Users don't have to
handle `a = 1` and `1 = a` any more. Does not expose catalyst.
Cons: Users have to deal with string parsing for complex projections.
Expressions pushed down are now are limited to those we decide to write
duplicate code for.
---
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 [email protected] or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
