Hoss Man created SOLR-8943:
------------------------------

             Summary: Solr tutorial can give errors when indexing techproduct 
data depending on what user has already indexed
                 Key: SOLR-8943
                 URL: https://issues.apache.org/jira/browse/SOLR-8943
             Project: Solr
          Issue Type: Bug
            Reporter: Hoss Man


http://lucene.apache.org/solr/quickstart.html currently says...

{quote}
Solr's install includes a handful of Solr XML formatted files with example data 
(mostly mocked tech product data). NOTE: This tech product data has a more 
domain-specific configuration, including schema and browse UI. The bin/solr 
script includes built-in support for this by running bin/solr start -e 
techproducts which not only starts Solr but also then indexes this data too (be 
sure to bin/solr stop -all before trying it out). However, the example below 
assumes Solr was started with bin/solr start -e cloud to stay consistent with 
all examples on this page, and thus the collection used is "gettingstarted", 
not "techproducts".
{quote}

But the assumption at the end of this paragraph -- that the sample techproduct 
documents can be added to the existing gettingstarted collection -- is 
fundementally flawed based on how the early sections of the tutorial are 
written.

With a completley clean, unused, empty gettingstarted demo collection, 
{{bin/post -c gettingstarted example/exampledocs/*.xml}} does work (on master 
anyway) but if you have been following along with teh tutorial you don't have a 
completley empty gettingstarted collection -- you will already have at a 
minimum the files in "docs", but also the user has been enocuraged to index 
arbitrary files from their computer...

{quote}
Indexing a directory of "rich" files
...
bin/post -c gettingstarted docs/
...
To index your own data, re-run the directory indexing command pointed to your 
own directory of documents. For example, on a Mac instead of docs/ try 
~/Documents/ or ~/Desktop/ ...
{quote}

...once the user has done that, it is not safe to assume that the techproduct 
sample documents can definitely be imported cleanly







--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org
For additional commands, e-mail: dev-h...@lucene.apache.org

Reply via email to