You are viewing a plain text version of this content. The canonical link for it is here.
Posted to solr-user@lucene.apache.org by madhav bahuguna <ma...@gmail.com> on 2014/07/11 09:04:58 UTC
delta import to detect deleted values
Hi,
How do i make delta import detect deleted values.
I do have a timestamp column to detect changes ,but the requirement is such
that rows will be deleted from the table.Every time i run delta import the
index still shows the deleted value.
How do i make delta import detect deleted values so that it does not show
deleted values.
--
Regards
Madhav Bahuguna
Re: delta import to detect deleted values
Posted by Chris Hostetter <ho...@fucit.org>.
: As Ahmet indicated, you must have a way to detect that deletions have
: happened. Marking rows as deleted with an active/inactive field is one
: way. Another way (the way that we use) is to have a delete trigger on
: the table that creates an entry in a delete tracking table.
If you have no control over the source datasource, and no way of knowing
if/when a document was deleted, your really only have 2 other options:
a) use some external system to track the "diff" of document uniqueKeys and
send deletes to solr based on that.
b) do a full rebuild everytime
for a lot of datasets, #b can acctually be quite fast and a lot less
headache then worrying about deletes -- if you keep a timestamp on every
doc, you can even do a "rolling rebuild" where you re-index all the live
docs, and then just remove any doc with a timestamp older then the start
of your re-index to prune the deletes.
-Hoss
http://www.lucidworks.com/
Re: delta import to detect deleted values
Posted by Shawn Heisey <so...@elyograg.org>.
On 7/11/2014 1:04 AM, madhav bahuguna wrote:
> How do i make delta import detect deleted values.
> I do have a timestamp column to detect changes ,but the requirement is such
> that rows will be deleted from the table.Every time i run delta import the
> index still shows the deleted value.
> How do i make delta import detect deleted values so that it does not show
> deleted values.
As Ahmet indicated, you must have a way to detect that deletions have
happened. Marking rows as deleted with an active/inactive field is one
way. Another way (the way that we use) is to have a delete trigger on
the table that creates an entry in a delete tracking table.
Either way, your indexing process will need customization to deal with
this additional information.
Thanks,
Shawn
Re: delta import to detect deleted values
Posted by Ahmet Arslan <io...@yahoo.com.INVALID>.
Hi,
You need a soft deletion column. Don't delete the entry instead mark it as deleted/inactive etc.
Ahmet
On Friday, July 11, 2014 10:12 AM, madhav bahuguna <ma...@gmail.com> wrote:
Hi,
How do i make delta import detect deleted values.
I do have a timestamp column to detect changes ,but the requirement is such
that rows will be deleted from the table.Every time i run delta import the
index still shows the deleted value.
How do i make delta import detect deleted values so that it does not show
deleted values.
--
Regards
Madhav Bahuguna