[CWB] corpus setup problem on CQPweb 3.0.5: cannot create
metadata , Error # 1062
Hardie, Andrew
a.hardie at lancaster.ac.uk
Fri May 18 15:34:04 CEST 2012
Hi Ray,
This is actually a bug - now fixed in SVN. Update to revision 316 and it should work. (you'll also get the new distribution table UI which I have just written).
You might need to run
delete from text_metadata_fieldswhere corpus="test";
before starting again to install the metadata.
Note also, your metadata input file should not contain header rows - just one row per text.
best
Andrew.
From: cwb-bounces at sslmit.unibo.it [mailto:cwb-bounces at sslmit.unibo.it] On Behalf Of ???
Sent: 18 May 2012 13:51
To: cwb at sslmit.unibo.it
Subject: [CWB] corpus setup problem on CQPweb 3.0.5: cannot create metadata , Error # 1062
Hi there,
This is Ray Wu from China. This is my first time being here. My solution to you all on the list.
I am a college ESL teacher and just a few days ago I started to learn CQPweb, both for teaching and research. Now I am fiddling with a toy corpus on CQPweb to get initiated.
My toy corpus is as suggested in Andrew's paper: test.vrt (encoded in ISO-8859-1)
<text id="test">
<s>
The AT0 the
cat NN1 cat
sat VVD sit
on PRP on
the AT0 the
mat NN1 mat
. PUN .
</s>
<s>
Many DT0 many
cats NN2 cat
sit VVB sit
on PRP on
mats NN2 mat
. PUN .
</s>
</text>
My metadata file: test_meta.dat (encoded in ISO-8859-1, tab separated)
text_id genre sampled
test press &n bsp; all
Presently, the corpus can be loaded into CQPweb without any fuss. But when I arrived at the "Admin tools for managing corpus metadata" page, I hit upon an error. Here are my then choices on that page:
field 1 genre classification
field 2 sampled classification
After clicking "install metadata table using the settings above", I got an error message like this:
A mySQL query did not run successfully!
Error # 1062: Duplicate entry 'test-__HANDLE' for key 1
I then peeped into MySQL and found the following 4 tables concerning metadata:
mysql> show tables;
...
corpus_metadata_fixed (empty)
corpus_metadata_variable (empty)
...
text_metadata_fields
text_metadata_values (empty)
...
mysql> select * from text_metadata_fields ;
+--------+----------+-------------+-------------------+
| corpus | handle | description | is_classification |
+--------+----------+-------------+-------------------+
| test | genre | | 1 |
| test | sampled | | 1 |
| test | __HANDLE | | 0 |
+--------+----------+-------------+-------------------+
But if I click "create mini malist metadata table", a metadata table can be created successfully and I could start to query.
mysql> select * from text_metadata_for_test;
+---------+-------+-----------+---------+
| text_id | words | cqp_begin | cqp_end |
+---------+-------+-----------+---------+
| test | 13 | 0 | 12 |
+---------+-------+-----------+---------+
This frustrates me as I know that without metadata, a corpus is of little value for search/research. Has anyone encountered similar messages before?
I have browsed all the archived mailinglist but found no direct answer to this problem (but I haven't looked at the source code yet). I don't know whether this indicates I need to manually add a few columns to the text_metadata_for_test table or I have just missed something important to get it done. Thanks for any pointers.
My thanks also goes to Andrew for a previous help regarding a CQPweb 3.0.5 file-write permission problem in a personal emai and pointing me at here. Thank you, Andrew.
PS: my computer paratemers:
System: Ubuntu 8.04
Apache: 2.0.63
MySQL: 5.0.88
PHP: 5.2.12 (lower than expected 5.3.0)
Perl: 5.8.8
CWB: 3.0.0
Linux utilites: awk, tar, gzip, iconv
Wu Liangping
School of International Studies
Hunan University of Commerce
PO Box 410000
Changsha, China
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://liste.sslmit.unibo.it/pipermail/cwb/attachments/20120518/6020b34f/attachment.htm
More information about the CWB
mailing list