Uploaded image for project: 'Pig'
  1. Pig
  2. PIG-2193

Using HBaseStorage to scan 2 tables in the same Map job produces bad data

    XMLWordPrintableJSON

Details

    • Bug
    • Status: Closed
    • Major
    • Resolution: Fixed
    • 0.8.1
    • 0.9.1, 0.10.0
    • None
    • None
    • HBase 0.90.3, Hadoop 0.20-append

    • Joining two HBase tables sometimes failed because mappers got confused about which of the tables they were supposed to scan.

    Description

      I've some data in HBase 0.90.3 and I run a simple script on them.

      This script badly returns 0 records. From time to time, under yet undefined conditions, the same script on the same data works (it return correct data).

      When data are loaded from HDFS instead of HBase, the script runs perfectly.

      Here is the script loading from HDFS (works):

      start_sessions = LOAD 'start_sessions' AS (sid:chararray, infoid:chararray, imei:chararray, start:long);
      end_sessions = LOAD 'end_sessions' AS (sid:chararray, end:long, locid:chararray);
      infos = LOAD 'infos' AS (infoid:chararray, network_type:chararray, network_subtype:chararray, locale:chararray, version_name:chararray, carrier_country:chararray, carrier_name:chararray, phone_manufacturer:chararray, phone_model:chararray, firmware_version:chararray, firmware_name:chararray);
      sessions = JOIN start_sessions BY sid, end_sessions BY sid;
      sessions = FILTER sessions BY end > start AND end - start < 86400000L;
      sessions = JOIN sessions BY infoid, infos BY infoid;
      sessions = LIMIT sessions 100;
      dump sessions;

      The same script loading from HBase (don't work):

      start_sessions = LOAD 'startSession' USING org.apache.pig.backend.hadoop.hbase.HBaseStorage('meta:sid meta:infoid meta:imei meta:timestamp') AS (sid:chararray, infoid:chararray, imei:chararray, start:long);
      end_sessions = LOAD 'endSession' USING org.apache.pig.backend.hadoop.hbase.HBaseStorage('meta:sid meta:timestamp meta:locid') AS (sid:chararray, end:long, locid:chararray);
      infos = LOAD 'info' USING org.apache.pig.backend.hadoop.hbase.HBaseStorage('meta:infoid data:networkType data:networkSubtype data:locale data:applicationVersionName data:carrierCountry data:carrierName data:phoneManufacturer data:phoneModel data:firmwareVersion data:firmwareName') AS (infoid:chararray, network_type:chararray, network_subtype:chararray, locale:chararray, version_name:chararray, carrier_country:chararray, carrier_name:chararray, phone_manufacturer:chararray, phone_model:chararray, firmware_version:chararray, firmware_name:chararray);
      sessions = JOIN start_sessions BY sid, end_sessions BY sid;
      sessions = FILTER sessions BY end > start AND end - start < 86400000L;
      sessions = JOIN sessions BY infoid, infos BY infoid;
      sessions = LIMIT sessions 100;
      dump sessions;

      I guess it definitively means there is a nasty bug in the HBase loader.

      Attachments

        1. PIG-2193.patch
          1 kB
          Raghu Angadi
        2. PIG-2193.patch
          9 kB
          Raghu Angadi

        Activity

          People

            rangadi Raghu Angadi
            vbarat Vincent BARAT
            Votes:
            0 Vote for this issue
            Watchers:
            5 Start watching this issue

            Dates

              Created:
              Updated:
              Resolved: