0

I am uploading many csv files

currency.csv file:

code,currency_name,currency_decimals
AUD,Australian Dollar,2
GBP,Pound Sterling,2
...
...

currency_holidays.csv file:

code,holiday_date,holiday_name
AUD,02/01/2012,New Year's Day Observed
AUD,26/01/2012,Australia Day
...
...

NOTE: uniqueKey is set to 'code' in solr configuration file

  1. if I upload these files in to solr single core it would overwrite the matching currency recordes e.g. AUD. Right?

  2. is it better to have core per file? i.e. multiple cores.

This is my previous post:

apache solr csv file same values

What is the best solution? I need help. Hope someone can help out.

Thanks GM

Community
  • 1
  • 1
user2279337
  • 691
  • 5
  • 13
  • 26
  • Please see answer to this previous question - http://stackoverflow.com/questions/15030176/solrj-indexing-multiple-classes-and-ensuring-document-uniqueness/15033183 – Paige Cook May 23 '13 at 16:49

1 Answers1

0

Some of the points you might want to think upon :-

  1. If you have completely different entities with nothing in common and not dependant as well (no joins), it would be better to have them as Separate Cores.
  2. This would be a much cleaner approach.
  3. As there might be fields which have a common name and would need to be analyzed in different ways as well as
  4. Search behaving in different ways for fields and their boost
  5. This would also be manageable if the data is huge.

However, if you have a very small dataset with none of the above concerning you just go with a single core.
You Unique keys you can prefix the ids with the type e.g. curreny_aud and holiday_aud which will help you keep the entities seperate and prevent overwriting.

Jayendra
  • 52,349
  • 4
  • 80
  • 90
  • Jayendra, thanks for the replay. For now I will be going with different cores. I did initially think about curreny_aud and holiday_aud which could be the alternative. Thanks again. – user2279337 May 24 '13 at 08:28