Setting the uprefix parameter of SolrCell (ERH) to something like "attr_"
will result in all metatdata attributes that are not named in the Solr
schema being given the "attr_" prefix to their metadata attribute names. For
example,
curl "http://localhost:8983/solr/update/extract?literal.id=doc-1\
&commit=true&uprefix=attr_" -F "my.pdf=@my.pdf"
Once you fixed out which of the metadata you want to keep, either add those
metadata attribute names to your schema, or
add explicit SolrCell field mappings for each piece of metadata:
&fmap.my-field=metadata-name.
-- Jack Krupansky
-----Original Message-----
From: Gian Maria Ricci
Sent: Monday, May 27, 2013 4:21 AM
To: solr-user@lucene.apache.org
Subject: RE: Tika: How can I import automatically all metadata without
specifiying them explicitly
Thanks for the help.
@Alexandre: Thanks for the suggestion, I'll try to use an
ExtractingRequestHandler, I thought that I was missing some DIH option :).
@Erik: I'm interested in knowing them all to do various form of analysis. I
have documents coming from heterogeneous sources and I'm interested in
searching inside the content, but also being able to extract all possible
metadata. I'm working in .Net so it is useful letting tika doing everything
for me directly in solr and then retrieve all metadata for matched
documents.
Thanks again to everyone.
--
Gian Maria Ricci
Mobile: +39 320 0136949
-----Original Message-----
From: Erick Erickson [mailto:erickerick...@gmail.com]
Sent: Sunday, May 26, 2013 5:30 PM
To: solr-user@lucene.apache.org; Gian Maria Ricci
Subject: Re: Tika: How can I import automatically all metadata without
specifiying them explicitly
In addition to Alexandre's comment:
bq: ...I'd like to import in my index all metadata
Be a little careful here, this isn't actually very useful in my experience.
Sure
it's nice to have all that data in the index, but... how do you search it
meaningfully?
Consider that some doc may have an "author" metadata field. Another may have
a "last editor" field. Yet another may have a "main author" field. If you
add all these as their field name, what do you do to search for "author"?
Somehow you have to create a mapping between the various metadata names and
something that's searchable, why not do this at index time?
Not to mention I've seen this done and the result may be literally hundreds
of different metadata fields which are not very useful.
All that said, it may be perfectly valid to inde them all, but before going
there it's worth considering whether the result is actually _useful_.
Best
Erick
On Sat, May 25, 2013 at 4:44 AM, Gian Maria Ricci
<alkamp...@nablasoft.com>wrote:
Hi to everyone,****
** **
I've configured import of a document folder with
FileListEntityProcessor, everything went smooth on the first try, but
I have a simple question. I'm able to map metadata without any
problem, but I'd like to import in my index all metadata, not only
those I've configured with field nodes. In this example I've imported
Author and title, but I does not know in advance which metadata a
document could have and I wish to have all of them inside my
index.****
** **
Here is my import config. It is the first try with importing with tika
and probably I'm missing a simple stuff.****
** **
<dataConfig> ****
<dataSource type="BinFileDataSource" />****
<document>****
<entity name="files"
dataSource="null" rootEntity="false"****
processor="FileListEntityProcessor" ****
baseDir="c:/temp/docs"
fileName=".*\.(doc)|(pdf)|(docx)"****
onError="skip"****
recursive="true">****
<field
column="file" name="id" />****
<field
column="fileAbsolutePath" name="path" />****
<field
column="fileSize" name="size" />****
<field
column="fileLastModified" name="lastModified" />****
****
<entity **
**
name="documentImport" ****
processor="TikaEntityProcessor"****
url="${files.fileAbsolutePath}" ****
format="text">****
<field column="file" name="fileName"/>****
<field column="Author" name="author" meta="true"/>****
<field column="title" name="title" meta="true"/>****
<field column="text" name="text"/>****
</entity>*
***
</entity>****
</document> ****
</dataConfig> ****
** **
** **
--****
Gian Maria Ricci****
Mobile: +39 320 0136949****
<http://mvp.microsoft.com/en-us/mvp/Gian%20Maria%20Ricci-4025635> [image:
https://encrypted-tbn1.gstatic.com/images?q=tbn:ANd9GcQyg0wiW_QuTxl-rn
uVR2P0jGuj4qO3I9attctCNarL--FC3vdPYg]<http://www.linkedin.com/in/gianm
ariaricci>
[image:
https://encrypted-tbn2.gstatic.com/images?q=tbn:ANd9GcT8z0HpwpDSjDWw1I
59Yx7HmF79u-NnP0NYeYYyEyWM1WtIbOl7]<https://twitter.com/alkampfer>
[image:
https://encrypted-tbn1.gstatic.com/images?q=tbn:ANd9GcQQWMj687BGGypKMU
Tub_lkUrull1uU2LTx0K2tDBeu3mNUr7Oxlg]<http://feeds.feedburner.com/Alka
mpferEng>
[image:
https://encrypted-tbn3.gstatic.com/images?q=tbn:ANd9GcSkTG_lPTPFe470xf
DtiInUtseqKcuV_lvI5h_-8t_3PsY5ikg3]
****
** **
** **