[
https://issues.apache.org/jira/browse/HBASE-16439?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Jean-Marc Spaggiari resolved HBASE-16439.
-----------------------------------------
Resolution: Not A Problem
Not an Apache HBase 0.98 issue. Fixed on Apache side. Might be vendor specific.
Please contact Cloudera support or ask on the Cloudera Community forum.
> Row mutation: missing rows with Put contains thousands of columns
> -----------------------------------------------------------------
>
> Key: HBASE-16439
> URL: https://issues.apache.org/jira/browse/HBASE-16439
> Project: HBase
> Issue Type: Bug
> Components: hbase
> Affects Versions: 0.98.6
> Environment: Cloudera CDH 5.3.3
> Reporter: Stas Makutin
> Attachments: HBaseAtomic.zip
>
>
> Some rows is missing when put them into a table using mutateRow method and
> each row has thousands of columns. The code excerpt:
> {code}
> try (HConnection hc = HConnectionManager.createConnection(conf)) {
> try (HTableInterface table = hc.getTable(tableName)) {
> final LocalDate startDate = LocalDate.of(1980, 01, 01);
> for (int i = 0; i < 15000; i++) {
> byte[] row = Bytes.toBytes(Integer.toString(i));
> long ts = System.currentTimeMillis();
> Put put = new Put(row, ts + 1);
> LocalDate date = startDate;
> for (int j = 0; j < 5000; j++) {
> put.add(
> family,
>
> Bytes.toBytes(DateTimeFormatter.BASIC_ISO_DATE.format(date)),
> Bytes.toBytes(Integer.toString(j))
> );
> date = date.plusDays(1);
> }
>
> RowMutations rm = new RowMutations(row);
> rm.add(put);
>
> table.mutateRow(rm);
> }
> }
> }
> {code}
> Resulting number of rows varies - sometimes it 200, sometimes it 8000. But
> never expected 15000.
> The full code of test application is attached.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
