Delete row that has columns with future timestamp
Hello
I use HBase version 0.98.9-hadoop1 with Hadoop version 1.2.1 . when i
delete row that has columns with future timestamp, delete not affect and
row still surviving.
For example when i put a row with future timestamp:
Put p = new Put(Bytes.toBytes("key1"));
p.add(Bytes.toBytes("C"), Bytes.toBytes("q1"), 2L,
Bytes.toBytes("test-val"));
table.put(p);
After put, when i scan my table, the result is:
ROW COLUMN+CELL
key1column=C:q1, timestamp=2, value=test-val
When i delete this row with following code:
Delete d = new Delete(Bytes.toBytes("key1"));
table.delete(d);
OR with this code:
Delete d = new Delete(Bytes.toBytes("key1"), Long.MAX_VALUE);
table.delete(d);
After each two deletes the result of scan is:
ROW COLUMN+CELL
key1column=C:q1, timestamp=2, value=test-val
And raw scan result is:
ROW COLUMN+CELL
key1column=C:, timestamp=1466931500501, type=DeleteFamily
key1column=C:q1, timestamp=2, value=test-val
But when i change the timestamp of delete to Long.MAX_VALUE-1, this delete
works. Can anyone help me with this?
Re: Delete row that has columns with future timestamp
you need to run a major compact after deletion
lise
> On 26 Jun 2016, at 11:20, M. BagherEsmaeily wrote:
>
> Hello
> I use HBase version 0.98.9-hadoop1 with Hadoop version 1.2.1 . when i
> delete row that has columns with future timestamp, delete not affect and
> row still surviving.
>
> For example when i put a row with future timestamp:
> Put p = new Put(Bytes.toBytes("key1"));
> p.add(Bytes.toBytes("C"), Bytes.toBytes("q1"), 2L,
> Bytes.toBytes("test-val"));
> table.put(p);
>
> After put, when i scan my table, the result is:
> ROW COLUMN+CELL
> key1column=C:q1, timestamp=2, value=test-val
>
> When i delete this row with following code:
> Delete d = new Delete(Bytes.toBytes("key1"));
> table.delete(d);
>
> OR with this code:
> Delete d = new Delete(Bytes.toBytes("key1"), Long.MAX_VALUE);
> table.delete(d);
>
> After each two deletes the result of scan is:
> ROW COLUMN+CELL
> key1column=C:q1, timestamp=2, value=test-val
>
> And raw scan result is:
> ROW COLUMN+CELL
> key1column=C:, timestamp=1466931500501, type=DeleteFamily
> key1column=C:q1, timestamp=2, value=test-val
>
>
> But when i change the timestamp of delete to Long.MAX_VALUE-1, this delete
> works. Can anyone help me with this?
Re: Delete row that has columns with future timestamp
these problem doesn't solve with major compact!! Assuming the problem is
solved with major compact, in this case, it's still a bug.
On Sun, Jun 26, 2016 at 3:08 PM, Lise Regnier
wrote:
> you need to run a major compact after deletion
> lise
>
> > On 26 Jun 2016, at 11:20, M. BagherEsmaeily
> wrote:
> >
> > Hello
> > I use HBase version 0.98.9-hadoop1 with Hadoop version 1.2.1 . when i
> > delete row that has columns with future timestamp, delete not affect and
> > row still surviving.
> >
> > For example when i put a row with future timestamp:
> > Put p = new Put(Bytes.toBytes("key1"));
> > p.add(Bytes.toBytes("C"), Bytes.toBytes("q1"), 2L,
> > Bytes.toBytes("test-val"));
> > table.put(p);
> >
> > After put, when i scan my table, the result is:
> > ROW COLUMN+CELL
> > key1column=C:q1, timestamp=2, value=test-val
> >
> > When i delete this row with following code:
> > Delete d = new Delete(Bytes.toBytes("key1"));
> > table.delete(d);
> >
> > OR with this code:
> > Delete d = new Delete(Bytes.toBytes("key1"), Long.MAX_VALUE);
> > table.delete(d);
> >
> > After each two deletes the result of scan is:
> > ROW COLUMN+CELL
> > key1column=C:q1, timestamp=2, value=test-val
> >
> > And raw scan result is:
> > ROW COLUMN+CELL
> > key1column=C:, timestamp=1466931500501, type=DeleteFamily
> > key1column=C:q1, timestamp=2, value=test-val
> >
> >
> > But when i change the timestamp of delete to Long.MAX_VALUE-1, this
> delete
> > works. Can anyone help me with this?
>
>
Re: Delete row that has columns with future timestamp
Hi,
This is a known issue and I think it is solved is more recent versions. Do
you have the option to upgrade?
JMS
Le 2016-06-26 07:00, "M. BagherEsmaeily" a écrit :
> these problem doesn't solve with major compact!! Assuming the problem is
> solved with major compact, in this case, it's still a bug.
>
> On Sun, Jun 26, 2016 at 3:08 PM, Lise Regnier
> wrote:
>
> > you need to run a major compact after deletion
> > lise
> >
> > > On 26 Jun 2016, at 11:20, M. BagherEsmaeily
> > wrote:
> > >
> > > Hello
> > > I use HBase version 0.98.9-hadoop1 with Hadoop version 1.2.1 . when i
> > > delete row that has columns with future timestamp, delete not affect
> and
> > > row still surviving.
> > >
> > > For example when i put a row with future timestamp:
> > > Put p = new Put(Bytes.toBytes("key1"));
> > > p.add(Bytes.toBytes("C"), Bytes.toBytes("q1"), 2L,
> > > Bytes.toBytes("test-val"));
> > > table.put(p);
> > >
> > > After put, when i scan my table, the result is:
> > > ROW COLUMN+CELL
> > > key1column=C:q1, timestamp=2, value=test-val
> > >
> > > When i delete this row with following code:
> > > Delete d = new Delete(Bytes.toBytes("key1"));
> > > table.delete(d);
> > >
> > > OR with this code:
> > > Delete d = new Delete(Bytes.toBytes("key1"), Long.MAX_VALUE);
> > > table.delete(d);
> > >
> > > After each two deletes the result of scan is:
> > > ROW COLUMN+CELL
> > > key1column=C:q1, timestamp=2, value=test-val
> > >
> > > And raw scan result is:
> > > ROW COLUMN+CELL
> > > key1column=C:, timestamp=1466931500501, type=DeleteFamily
> > > key1column=C:q1, timestamp=2, value=test-val
> > >
> > >
> > > But when i change the timestamp of delete to Long.MAX_VALUE-1, this
> > delete
> > > works. Can anyone help me with this?
> >
> >
>
Re: Delete row that has columns with future timestamp
Hey M.,
Just to follow up on what JMS said, this was fixed in April 2014 (details
at https://issues.apache.org/jira/browse/HBASE-10118), so running a version
of HBase in which the patch went in is probably your best option.
-Dima
On Sunday, June 26, 2016, Jean-Marc Spaggiari
wrote:
> Hi,
>
> This is a known issue and I think it is solved is more recent versions. Do
> you have the option to upgrade?
>
> JMS
> Le 2016-06-26 07:00, "M. BagherEsmaeily" > a écrit :
>
> > these problem doesn't solve with major compact!! Assuming the problem is
> > solved with major compact, in this case, it's still a bug.
> >
> > On Sun, Jun 26, 2016 at 3:08 PM, Lise Regnier >
> > wrote:
> >
> > > you need to run a major compact after deletion
> > > lise
> > >
> > > > On 26 Jun 2016, at 11:20, M. BagherEsmaeily >
> > > wrote:
> > > >
> > > > Hello
> > > > I use HBase version 0.98.9-hadoop1 with Hadoop version 1.2.1 . when i
> > > > delete row that has columns with future timestamp, delete not affect
> > and
> > > > row still surviving.
> > > >
> > > > For example when i put a row with future timestamp:
> > > > Put p = new Put(Bytes.toBytes("key1"));
> > > > p.add(Bytes.toBytes("C"), Bytes.toBytes("q1"), 2L,
> > > > Bytes.toBytes("test-val"));
> > > > table.put(p);
> > > >
> > > > After put, when i scan my table, the result is:
> > > > ROW COLUMN+CELL
> > > > key1column=C:q1, timestamp=2, value=test-val
> > > >
> > > > When i delete this row with following code:
> > > > Delete d = new Delete(Bytes.toBytes("key1"));
> > > > table.delete(d);
> > > >
> > > > OR with this code:
> > > > Delete d = new Delete(Bytes.toBytes("key1"), Long.MAX_VALUE);
> > > > table.delete(d);
> > > >
> > > > After each two deletes the result of scan is:
> > > > ROW COLUMN+CELL
> > > > key1column=C:q1, timestamp=2, value=test-val
> > > >
> > > > And raw scan result is:
> > > > ROW COLUMN+CELL
> > > > key1column=C:, timestamp=1466931500501, type=DeleteFamily
> > > > key1column=C:q1, timestamp=2, value=test-val
> > > >
> > > >
> > > > But when i change the timestamp of delete to Long.MAX_VALUE-1, this
> > > delete
> > > > works. Can anyone help me with this?
> > >
> > >
> >
>
Re: Delete row that has columns with future timestamp
HBASE-10118 was integrated into 0.98.2
The user was running 0.98.9
Hmm
On Sun, Jun 26, 2016 at 12:05 PM, Dima Spivak wrote:
> Hey M.,
>
> Just to follow up on what JMS said, this was fixed in April 2014 (details
> at https://issues.apache.org/jira/browse/HBASE-10118), so running a
> version
> of HBase in which the patch went in is probably your best option.
>
> -Dima
>
> On Sunday, June 26, 2016, Jean-Marc Spaggiari
> wrote:
>
> > Hi,
> >
> > This is a known issue and I think it is solved is more recent versions.
> Do
> > you have the option to upgrade?
> >
> > JMS
> > Le 2016-06-26 07:00, "M. BagherEsmaeily" > > a écrit :
> >
> > > these problem doesn't solve with major compact!! Assuming the problem
> is
> > > solved with major compact, in this case, it's still a bug.
> > >
> > > On Sun, Jun 26, 2016 at 3:08 PM, Lise Regnier > >
> > > wrote:
> > >
> > > > you need to run a major compact after deletion
> > > > lise
> > > >
> > > > > On 26 Jun 2016, at 11:20, M. BagherEsmaeily > >
> > > > wrote:
> > > > >
> > > > > Hello
> > > > > I use HBase version 0.98.9-hadoop1 with Hadoop version 1.2.1 .
> when i
> > > > > delete row that has columns with future timestamp, delete not
> affect
> > > and
> > > > > row still surviving.
> > > > >
> > > > > For example when i put a row with future timestamp:
> > > > > Put p = new Put(Bytes.toBytes("key1"));
> > > > > p.add(Bytes.toBytes("C"), Bytes.toBytes("q1"), 2L,
> > > > > Bytes.toBytes("test-val"));
> > > > > table.put(p);
> > > > >
> > > > > After put, when i scan my table, the result is:
> > > > > ROW COLUMN+CELL
> > > > > key1column=C:q1, timestamp=2, value=test-val
> > > > >
> > > > > When i delete this row with following code:
> > > > > Delete d = new Delete(Bytes.toBytes("key1"));
> > > > > table.delete(d);
> > > > >
> > > > > OR with this code:
> > > > > Delete d = new Delete(Bytes.toBytes("key1"), Long.MAX_VALUE);
> > > > > table.delete(d);
> > > > >
> > > > > After each two deletes the result of scan is:
> > > > > ROW COLUMN+CELL
> > > > > key1column=C:q1, timestamp=2, value=test-val
> > > > >
> > > > > And raw scan result is:
> > > > > ROW COLUMN+CELL
> > > > > key1column=C:, timestamp=1466931500501, type=DeleteFamily
> > > > > key1column=C:q1, timestamp=2, value=test-val
> > > > >
> > > > >
> > > > > But when i change the timestamp of delete to Long.MAX_VALUE-1, this
> > > > delete
> > > > > works. Can anyone help me with this?
> > > >
> > > >
> > >
> >
>
Re: Delete row that has columns with future timestamp
Yes, Major Compact is not my problem that is mentioned in the HBASE-10118,
My problem is surviving row after delete even logically, And the row is
still in the scanner result.
I think the problem relates to timestamp of Delete object. When i set
it to Long.MAX_VALUE
(if you set no timestamp in Delete class constructor, Long.MAX_VALUE was
set as default value), these problem happens and according to the raw
scan output,
timestamp of delete record is time that delete occured instead of
Long.MAX_VALUE
value.
Raw scan result:
ROW COLUMN+CELL
key1column=C:, timestamp=*1466931500501*, type=*DeleteFamily*
key1column=C:q1, timestamp=2, value=test-val
On Mon, Jun 27, 2016 at 5:57 PM, Ted Yu wrote:
> HBASE-10118 was integrated into 0.98.2
>
> The user was running 0.98.9
>
> Hmm
>
> On Sun, Jun 26, 2016 at 12:05 PM, Dima Spivak
> wrote:
>
> > Hey M.,
> >
> > Just to follow up on what JMS said, this was fixed in April 2014 (details
> > at https://issues.apache.org/jira/browse/HBASE-10118), so running a
> > version
> > of HBase in which the patch went in is probably your best option.
> >
> > -Dima
> >
> > On Sunday, June 26, 2016, Jean-Marc Spaggiari
> > wrote:
> >
> > > Hi,
> > >
> > > This is a known issue and I think it is solved is more recent versions.
> > Do
> > > you have the option to upgrade?
> > >
> > > JMS
> > > Le 2016-06-26 07:00, "M. BagherEsmaeily" > > > a écrit :
> > >
> > > > these problem doesn't solve with major compact!! Assuming the problem
> > is
> > > > solved with major compact, in this case, it's still a bug.
> > > >
> > > > On Sun, Jun 26, 2016 at 3:08 PM, Lise Regnier <
> lise.regn...@gmail.com
> > > >
> > > > wrote:
> > > >
> > > > > you need to run a major compact after deletion
> > > > > lise
> > > > >
> > > > > > On 26 Jun 2016, at 11:20, M. BagherEsmaeily <
> mbesmae...@gmail.com
> > > >
> > > > > wrote:
> > > > > >
> > > > > > Hello
> > > > > > I use HBase version 0.98.9-hadoop1 with Hadoop version 1.2.1 .
> > when i
> > > > > > delete row that has columns with future timestamp, delete not
> > affect
> > > > and
> > > > > > row still surviving.
> > > > > >
> > > > > > For example when i put a row with future timestamp:
> > > > > > Put p = new Put(Bytes.toBytes("key1"));
> > > > > > p.add(Bytes.toBytes("C"), Bytes.toBytes("q1"), 2L,
> > > > > > Bytes.toBytes("test-val"));
> > > > > > table.put(p);
> > > > > >
> > > > > > After put, when i scan my table, the result is:
> > > > > > ROW COLUMN+CELL
> > > > > > key1column=C:q1, timestamp=2, value=test-val
> > > > > >
> > > > > > When i delete this row with following code:
> > > > > > Delete d = new Delete(Bytes.toBytes("key1"));
> > > > > > table.delete(d);
> > > > > >
> > > > > > OR with this code:
> > > > > > Delete d = new Delete(Bytes.toBytes("key1"), Long.MAX_VALUE);
> > > > > > table.delete(d);
> > > > > >
> > > > > > After each two deletes the result of scan is:
> > > > > > ROW COLUMN+CELL
> > > > > > key1column=C:q1, timestamp=2, value=test-val
> > > > > >
> > > > > > And raw scan result is:
> > > > > > ROW COLUMN+CELL
> > > > > > key1column=C:, timestamp=1466931500501, type=DeleteFamily
> > > > > > key1column=C:q1, timestamp=2, value=test-val
> > > > > >
> > > > > >
> > > > > > But when i change the timestamp of delete to Long.MAX_VALUE-1,
> this
> > > > > delete
> > > > > > works. Can anyone help me with this?
> > > > >
> > > > >
> > > >
> > >
> >
>
