[SPARK-32251][SQL][DOCS][TESTS] Fix SQL keyword document #29055
There are no files selected for viewing
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,172 @@ | ||
| /* | ||
| * 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.catalyst | ||
|
|
||
| import java.io.File | ||
| import java.nio.file.Files | ||
|
|
||
| import scala.collection.JavaConverters._ | ||
| import scala.collection.mutable | ||
|
|
||
| import org.apache.spark.SparkFunSuite | ||
| import org.apache.spark.sql.catalyst.plans.SQLHelper | ||
| import org.apache.spark.sql.catalyst.util.fileToString | ||
|
|
||
| trait SQLKeywordUtils extends SQLHelper { | ||
|
|
||
| val sqlSyntaxDefs = { | ||
| val sqlBasePath = { | ||
| java.nio.file.Paths.get(sparkHome, "sql", "catalyst", "src", "main", "antlr4", "org", | ||
| "apache", "spark", "sql", "catalyst", "parser", "SqlBase.g4").toFile | ||
| } | ||
| fileToString(sqlBasePath).split("\n") | ||
| } | ||
|
|
||
| // each element is an array of 4 string: the keyword name, reserve or not in Spark ANSI mode, | ||
| // Spark non-ANSI mode, and the SQL standard. | ||
| val keywordsInDoc: Array[Array[String]] = { | ||
| val docPath = { | ||
| java.nio.file.Paths.get(sparkHome, "docs", "sql-ref-ansi-compliance.md").toFile | ||
| } | ||
| fileToString(docPath).split("\n") | ||
| .dropWhile(!_.startsWith("|Keyword|")).drop(2).takeWhile(_.startsWith("|")) | ||
| .map(_.stripPrefix("|").split("\\|").map(_.trim)) | ||
| } | ||
|
|
||
| private def parseAntlrGrammars[T](startTag: String, endTag: String) | ||
| (f: PartialFunction[String, Seq[T]]): Set[T] = { | ||
| val keywords = new mutable.ArrayBuffer[T] | ||
| val default = (_: String) => Nil | ||
| var startTagFound = false | ||
| var parseFinished = false | ||
| val lineIter = sqlSyntaxDefs.toIterator | ||
| while (!parseFinished && lineIter.hasNext) { | ||
| val line = lineIter.next() | ||
| if (line.trim.startsWith(startTag)) { | ||
| startTagFound = true | ||
| } else if (line.trim.startsWith(endTag)) { | ||
| parseFinished = true | ||
| } else if (startTagFound) { | ||
| f.applyOrElse(line, default).foreach { symbol => | ||
| keywords += symbol | ||
| } | ||
| } | ||
| } | ||
| assert(keywords.nonEmpty && startTagFound && parseFinished, "cannot extract keywords from " + | ||
| s"the `SqlBase.g4` file, so please check if the start/end tags (`$startTag` and `$endTag`) " + | ||
| "are placed correctly in the file.") | ||
| keywords.toSet | ||
| } | ||
|
|
||
| // If a symbol does not have the same string with its literal (e.g., `SETMINUS: 'MINUS';`), | ||
| // we need to map a symbol to actual literal strings. | ||
| val symbolsToExpandIntoDifferentLiterals = { | ||
| val kwDef = """([A-Z_]+):(.+);""".r | ||
| val keywords = parseAntlrGrammars( | ||
| "//--SPARK-KEYWORD-LIST-START", "//--SPARK-KEYWORD-LIST-END") { | ||
| case kwDef(symbol, literalDef) => | ||
| val splitDefs = literalDef.split("""\|""") | ||
| val hasMultipleLiterals = splitDefs.length > 1 | ||
| // The case where a symbol has multiple literal definitions, | ||
| // e.g., `DATABASES: 'DATABASES' | 'SCHEMAS';`. | ||
| if (hasMultipleLiterals) { | ||
| // Filters out inappropriate entries, e.g., `!` in `NOT: 'NOT' | '!';` | ||
| val litDef = """([A-Z_]+)""".r | ||
| val literals = splitDefs.map(_.replaceAll("'", "").trim).toSeq.flatMap { | ||
| case litDef(lit) => Some(lit) | ||
| case _ => None | ||
| } | ||
| (symbol, literals) :: Nil | ||
| } else { | ||
| val literal = literalDef.replaceAll("'", "").trim | ||
| // The case where a symbol string and its literal string are different, | ||
| // e.g., `SETMINUS: 'MINUS';`. | ||
| if (symbol != literal) { | ||
| (symbol, literal :: Nil) :: Nil | ||
| } else { | ||
| Nil | ||
| } | ||
| } | ||
| } | ||
| keywords.toMap | ||
| } | ||
|
|
||
| // All the SQL keywords defined in `SqlBase.g4` | ||
| val allCandidateKeywords: Set[String] = { | ||
| val kwDef = """([A-Z_]+):.+;""".r | ||
| val keywords = parseAntlrGrammars( | ||
| "//--SPARK-KEYWORD-LIST-START", "//--SPARK-KEYWORD-LIST-END") { | ||
| // Parses a pattern, e.g., `AFTER: 'AFTER';` | ||
| case kwDef(symbol) => | ||
| if (symbolsToExpandIntoDifferentLiterals.contains(symbol)) { | ||
| symbolsToExpandIntoDifferentLiterals(symbol) | ||
| } else { | ||
| symbol :: Nil | ||
| } | ||
| } | ||
| // DIV is an operator but is also a keyword | ||
| keywords + "DIV" | ||
|
Member
There was a problem hiding this comment. @cloud-fan . Instead of adding exceptions here and line 136, can we switch the original lines in |
||
| } | ||
|
|
||
| val nonReservedKeywordsInAnsiMode: Set[String] = { | ||
| val kwDef = """\s*[\|:]\s*([A-Z_]+)\s*""".r | ||
| parseAntlrGrammars("//--ANSI-NON-RESERVED-START", "//--ANSI-NON-RESERVED-END") { | ||
| // Parses a pattern, e.g., ` | AFTER` | ||
| case kwDef(symbol) => | ||
| if (symbolsToExpandIntoDifferentLiterals.contains(symbol)) { | ||
| symbolsToExpandIntoDifferentLiterals(symbol) | ||
| } else { | ||
| symbol :: Nil | ||
| } | ||
| } + "DIV" // DIV is non-reserved | ||
| } | ||
|
|
||
| val reservedKeywordsInAnsiMode = allCandidateKeywords -- nonReservedKeywordsInAnsiMode | ||
| } | ||
|
|
||
| class SQLKeywordSuite extends SparkFunSuite with SQLKeywordUtils { | ||
| test("all keywords are documented") { | ||
| val documentedKeywords = keywordsInDoc.map(_.head).toSet | ||
| if (allCandidateKeywords != documentedKeywords) { | ||
| val undocumented = (allCandidateKeywords -- documentedKeywords).toSeq.sorted | ||
| fail("Some keywords are not documented: " + undocumented.mkString(", ")) | ||
| } | ||
| } | ||
|
|
||
| test("Spark keywords are documented correctly") { | ||
| val reservedKeywordsInDoc = keywordsInDoc.filter(_.apply(1) == "reserved").map(_.head).toSet | ||
| if (reservedKeywordsInAnsiMode != reservedKeywordsInDoc) { | ||
| val misImplemented = (reservedKeywordsInDoc -- reservedKeywordsInAnsiMode).toSeq.sorted | ||
| fail("Some keywords are documented as reserved but actually not: " + | ||
| misImplemented.mkString(", ")) | ||
| } | ||
| } | ||
|
|
||
| test("SQL 2016 keywords are documented correctly") { | ||
| withTempDir { dir => | ||
| val tmpFile = new File(dir, "tmp") | ||
| val is = Thread.currentThread().getContextClassLoader | ||
| .getResourceAsStream("ansi-sql-2016-reserved-keywords.txt") | ||
| Files.copy(is, tmpFile.toPath) | ||
| val reservedKeywordsInSql2016 = Files.readAllLines(tmpFile.toPath) | ||
| .asScala.filterNot(_.startsWith("--")).map(_.trim).toSet | ||
| val documented = keywordsInDoc.filter(_.last == "reserved").map(_.head).toSet | ||
| assert((documented -- reservedKeywordsInSql2016).isEmpty) | ||
| } | ||
| } | ||
| } | ||
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Is there any consensus to decide which version of ANSI SQL standard that we actually try to comply with?
Or any protocol to define the compatibility between Spark versions and ANSI SQL standards？
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I think the latest ANSI SQL standard is preferred.
Also to note that: ANSI compliant doesn't mean to follow everything in ANSI SQL. e.g. a keyword reserved in ANSI SQL can be non-reserved in Spark.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Looking at the history of this standard, the trend of it seems to add more and more keywords and change the non-reserved to reserved. I guess it's ok to follow the latest one here as literally we will add migration guide for every newly added keyword.