On 8/7/06, James Tu <[EMAIL PROTECTED]> wrote:
If I create a table for each user (I can potentially have hundreds of thousands of users), will MySQL be able to handle this? If I just have one table, I could potentially have millions of records in one table. Will MySQL be able to handle this?
Hi James, There are really two elements to this problem. The first element is how quickly MySQL can extract the messages you want from a large table. This requires that you know in advance the type of queries you're going to do (all messages by one user? all messages in a certain time window?) and be sure that these queries are approximately O(log N) rather than O(N) or worse. You will need to change your database design to fit the queries that you'll be doing. O(log N) queries would generally be characterized by the fields you're searching or sorting on being key fields (i.e. MySQL makes an index or BTREE or whatever it makes rather than having to go through the entire table linearly). The second element is data presentation. In developing web applications at least, if the first data on a page is displayed by the browser while the rest of the data is loading, the user perceives the load as being faster than it really is because the user is looking at the first data while the rest is loading. So, to make things more "snappy", you might do more than one query to avoid large result sets. Dave.
