Home/ Questions/Q 309155
In Process

The Archive Base Latest Questions

Editorial Team
  • 0
Asked: 2026-05-12T07:42:58+00:00

I have a dataset which contains a string key field and up to 50

  • 0

I have a dataset which contains a string key field and up to 50 keywords associated with that information. Once the data has been inserted into the database there will be very few writes (INSERTS) but mostly queries for one or more keywords.

I have read “Tagsystems: performance tests” which is MySQL based and it seems 2NF appears to be a good method for implementing this, however I was wondering if anyone had experience with doing this with SQL Server 2008 and very large datasets.

I am likely to initially have 1 million key fields which could have up to 50 keywords each.

Would a structure of

keyfield, keyword1, keyword2, ... , keyword50

be the best solution or two tables

keyid
keyfield
| 1
|
| M
keyid
keyword

Be a better idea if my queries are mostly going to be looking for results that have one or more keywords?

Share

Leave an answer

You must login to add an answer.

Need An Account,

1 Answer

  1. Editorial Team

    I would normalize a step further.

    You should have a table of unique KeyWords with an integer primary key column. Then, another association table that has KeyField and KeyWordId.

    KeyWords
----------
KeyWordId Int Identity(1,1)
KeyWord VarChar(200)

KeyFieldKeyWords
----------------
Keyfield Int
KeyWordId Int

    With 1 million keyfields having 50 keywords each, that’s 50 million rows. There will be a HUGE difference in performance if you have a table with 2 columns, each being an integer.

    • 0