hive-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "ASF GitHub Bot (Jira)" <j...@apache.org>
Subject [jira] [Work logged] (HIVE-21924) Split text files even if header/footer exists
Date Wed, 09 Oct 2019 16:03:00 GMT

     [ https://issues.apache.org/jira/browse/HIVE-21924?focusedWorklogId=325786&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-325786
]

ASF GitHub Bot logged work on HIVE-21924:
-----------------------------------------

                Author: ASF GitHub Bot
            Created on: 09/Oct/19 16:02
            Start Date: 09/Oct/19 16:02
    Worklog Time Spent: 10m 
      Work Description: sankarh commented on pull request #791: HIVE-21924
URL: https://github.com/apache/hive/pull/791#discussion_r333100168
 
 

 ##########
 File path: ql/src/java/org/apache/hadoop/hive/ql/io/SkippingTextInputFormat.java
 ##########
 @@ -0,0 +1,212 @@
+/*
+ * 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.hadoop.hive.ql.io;
+
+import org.apache.hadoop.fs.FSDataInputStream;
+import org.apache.hadoop.fs.FileSystem;
+import org.apache.hadoop.fs.Path;
+import org.apache.hadoop.mapred.FileSplit;
+import org.apache.hadoop.mapred.JobConf;
+import org.apache.hadoop.mapred.TextInputFormat;
+
+import java.io.IOException;
+import java.util.ArrayDeque;
+import java.util.Map;
+import java.util.Queue;
+import java.util.concurrent.ConcurrentHashMap;
+
+/**
+ * SkippingInputFormat is a header/footer aware input format. It truncates
+ * splits identified by TextInputFormat. Header and footers are removed
+ * from the splits.
+ */
+public class SkippingTextInputFormat extends TextInputFormat {
+
+  private final Map<Path, Long> startIndexMap = new ConcurrentHashMap<Path, Long>();
+  private final Map<Path, Long> endIndexMap = new ConcurrentHashMap<Path, Long>();
+  private JobConf conf;
+  private int headerCount;
+  private int footerCount;
+
+  @Override
+  public void configure(JobConf conf) {
+    this.conf = conf;
+    super.configure(conf);
+  }
+
+  public void configure(JobConf conf, int headerCount, int footerCount) {
+    configure(conf);
+    this.headerCount = headerCount;
+    this.footerCount = footerCount;
+  }
+
+  @Override
+  protected FileSplit makeSplit(Path file, long start, long length, String[] hosts) {
+    return makeSplitInternal(file, start, length, hosts, null);
+  }
+
+  @Override
+  protected FileSplit makeSplit(Path file, long start, long length, String[] hosts, String[]
inMemoryHosts) {
+    return makeSplitInternal(file, start, length, hosts, inMemoryHosts);
+  }
+
+  private FileSplit makeSplitInternal(Path file, long start, long length, String[] hosts,
String[] inMemoryHosts) {
+    long cachedStart;
+    long cachedEnd;
+    try {
+      cachedStart = getCachedStartIndex(file);
+      cachedEnd = getCachedEndIndex(file);
+    } catch (IOException e) {
+      LOG.warn("Could not detect header/footer", e);
+      return new NullRowsInputFormat.DummyInputSplit(file);
+    }
+    if (cachedStart > start + length) {
+      return new NullRowsInputFormat.DummyInputSplit(file);
+    }
+    if (cachedStart > start) {
+      length = length - (cachedStart - start);
+      start = cachedStart;
+    }
+    if (cachedEnd < start) {
+      return new NullRowsInputFormat.DummyInputSplit(file);
+    }
+    if (cachedEnd < start + length) {
+      length = cachedEnd - start;
+    }
+    if (inMemoryHosts == null) {
+      return super.makeSplit(file, start, length, hosts);
+    } else {
+      return super.makeSplit(file, start, length, hosts, inMemoryHosts);
+    }
+  }
+
+  private long getCachedStartIndex(Path path) throws IOException {
+    if (headerCount == 0) {
+      return 0;
+    }
+    Long startIndexForFile = startIndexMap.get(path);
+    if (startIndexForFile == null) {
+      FileSystem fileSystem;
+      FSDataInputStream fis = null;
+      fileSystem = path.getFileSystem(conf);
+      try {
+        fis = fileSystem.open(path);
+        for (int j = 0; j < headerCount; j++) {
+          if (fis.readLine() == null) {
+            startIndexMap.put(path, Long.MAX_VALUE);
+            return Long.MAX_VALUE;
+          }
+        }
+        // back 1 byte because readers skip the entire first row if split start is not 0
+        startIndexForFile = fis.getPos() - 1;
+      } finally {
+        if (fis != null) {
+          fis.close();
+        }
+      }
+      startIndexMap.put(path, startIndexForFile);
+    }
+    return startIndexForFile;
+  }
+
+  private long getCachedEndIndex(Path path) throws IOException {
+    Long endIndexForFile = endIndexMap.get(path);
+    if (endIndexForFile == null) {
+      final long bufferSectionSize = 5 * 1024;
+      FileSystem fileSystem = path.getFileSystem(conf);
+      long endOfFile = fileSystem.getFileStatus(path).getLen();
+      if (footerCount == 0) {
+        endIndexForFile = endOfFile;
+      } else {
+        long bufferSectionEnd = endOfFile;
+        long bufferSectionStart = Math.max(0, bufferSectionEnd - bufferSectionSize);
 
 Review comment:
   Yes. I agree.
 
----------------------------------------------------------------
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.
 
For queries about this service, please contact Infrastructure at:
users@infra.apache.org


Issue Time Tracking
-------------------

    Worklog Id:     (was: 325786)
    Time Spent: 3h 20m  (was: 3h 10m)

> Split text files even if header/footer exists
> ---------------------------------------------
>
>                 Key: HIVE-21924
>                 URL: https://issues.apache.org/jira/browse/HIVE-21924
>             Project: Hive
>          Issue Type: Improvement
>          Components: File Formats
>    Affects Versions: 2.4.0, 4.0.0, 3.2.0
>            Reporter: Prasanth Jayachandran
>            Assignee: Mustafa Iman
>            Priority: Major
>              Labels: pull-request-available
>         Attachments: HIVE-21924.2.patch, HIVE-21924.3.patch, HIVE-21924.4.patch, HIVE-21924.5.patch,
HIVE-21924.6.patch, HIVE-21924.patch
>
>          Time Spent: 3h 20m
>  Remaining Estimate: 0h
>
> https://github.com/apache/hive/blob/967a1cc98beede8e6568ce750ebeb6e0d048b8ea/ql/src/java/org/apache/hadoop/hive/ql/io/HiveInputFormat.java#L494-L503

> {code}
>     int headerCount = 0;
>     int footerCount = 0;
>     if (table != null) {
>       headerCount = Utilities.getHeaderCount(table);
>       footerCount = Utilities.getFooterCount(table, conf);
>       if (headerCount != 0 || footerCount != 0) {
>         // Input file has header or footer, cannot be splitted.
>         HiveConf.setLongVar(conf, ConfVars.MAPREDMINSPLITSIZE, Long.MAX_VALUE);
>       }
>     }
> {code}
> this piece of code makes the CSV (or any text files with header/footer) files not splittable
if header or footer is present. 
> If only header is present, we can find the offset after first line break and use that
to split. Similarly for footer, may be read few KB's of data at the end and find the last
line break offset. Use that to determine the data range which can be used for splitting. Few
reads during split generation are cheaper than not splitting the file at all.  



--
This message was sent by Atlassian Jira
(v8.3.4#803005)

Mime
View raw message