Mike, #2 is not 100% correct. Only the columns in the "lookup" table (table2) need to be indexed. Why? The optimizer first gets a limited set of table1 and then checks if it can join this subset with a key from table2. So for the join condition only table2 needs to be indexed.
However, you are right that in most of the cases you also want to have a subset out of table one first. For that you also should use an index, but in most cases this will be a different index. Example: CREATE TABLE table1 ( A INT NOT NULL, B INT, C INT, PRIMARY KEY(A), UNIQUE INDEX uidx1(B) ); CREATE TABLE table2 ( C INT NOT NULL, D INT, PRIMARY KEY(C) ); SELECT table2.D FROM table1, table2 WHERE table1.B = 25 AND table1.C = table2.C The optimizer now would first find all rows from table1 having B = 25 by using the INDEX uidx1 and then join table 2 via the columns table1.C = table2.C using the index PRIMARY KEY from table2. So for the JOIN you only need to have table2 indexed, there is no need to put an index on the column C on table1. BTW you can always see what the optimizer plans to do by setting EXPLAIN right before SELECT (EXPLAIN SELECT ...) What the former post was more about is REFERENTIAL INTEGRITY, which is something that is not (yet) include in MySQL (at least for MyISAM, for InnoDB it is coming with 5.0 I think). Cheers /rudy -----Original Message----- From: Mike Hillyer [mailto:[EMAIL PROTECTED] Sent: woensdag 9 juli 2003 17:19 To: Adam Gerson; [EMAIL PROTECTED] Cc: [EMAIL PROTECTED]; [EMAIL PROTECTED] Subject: RE: Can mysql handle this load? Well that all depends. The real reason for a primary key is to prevent duplicates, therefore, if a combination of fields needs to be unique, then a multiple primary key makes sense, especially if other tables will reference the field combination (for example, detail items on an invoice where the invoice number will not be unique, and the detail number will not be unique, but the combination of the two will be). As for question #2, if two fields, one on each table, will be used to join tables together, both fields should be indexed. Regards, Mike Hillyer www.vbmysql.com > -----Original Message----- > From: Adam Gerson [mailto:[EMAIL PROTECTED] > Sent: Wednesday, July 09, 2003 9:09 AM > To: [EMAIL PROTECTED] > Cc: [EMAIL PROTECTED]; [EMAIL PROTECTED] > Subject: Re: Can mysql handle this load? > > > Is it better to set multiple primary keys or to set one key and index > the other columns? If I have a primary key as a field in > another table > should it also be set as a key? > > > > Adam > > On Wednesday, July 9, 2003, at 10:41 AM, > <[EMAIL PROTECTED]> wrote: > > > i think this should be no problem... > > > > i'd think of some table layout like this: > > date int PRIMARY > > student_id int PRIMARY > > status int > > extra_data what-you-want > > > > then you should get about 360,000 records per year. > > i saw people on this list reporting about millions of > records etc... > > and i guess they had a little greater tables than you > should get here. > > > > but why would you want to move any previous records to > another table > > all the time? just keep it in one table and back up anything older > > than 5 years or so. that keeps your table at, say 50 MB, > and you can > > run real-time queries anytime :) > > > > -yves > > > > > > -----Ursprüngliche Nachricht----- > > Von: "Adam Gerson" <[EMAIL PROTECTED]> > > An: <[EMAIL PROTECTED]> > > Cc: <[EMAIL PROTECTED]> > > Gesendet: Mittwoch, 9. Juli 2003 15:46 > > Betreff: Can mysql handle this load? > > > > > >> I am writing an attendance system in php for my school. We have a > >> little less then 1000 students. For every day of the > school year one > >> record will be entered into a table for each student representing > >> their > >> attendance status (present, absent, late, etc...). I also > have several > >> other supporting tables for relationships. When it comes > to reporting > >> and querying this DB I am worried that it will very quickly become > >> very > >> large and slow. Can mysql handle this? Are there any techniques to > >> speed it up? I will trying indexing major columns. > >> > >> I have also considered keeping all previous days attendance in a > >> separate table from the current days attendance and moving > things over > >> in the middle of the night. This way any operations on the current > >> days > >> data will go quickly, but reports on long term things will still be > >> slow. Good idea? > >> > >> Thanks, > >> Adam > >> > >> > >> > >> > >> > >> ------------------------------- > >> Adam Gerson > >> Systems Administrator / Computer Teacher > >> Columbia Grammar and Prep School > >> 212-749-6200 > >> [EMAIL PROTECTED] > >> www.cgps.org > >> > >> > >> -- > >> MySQL General Mailing List > >> For list archives: http://lists.mysql.com/mysql > >> To unsubscribe: > >> http://lists.mysql.com/[EMAIL PROTECTED] > >> > > > > > -- > MySQL General Mailing List > For list archives: http://lists.mysql.com/mysql > To unsubscribe: > http://lists.mysql.com/mysql?> [EMAIL PROTECTED] > > -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe: http://lists.mysql.com/[EMAIL PROTECTED] -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe: http://lists.mysql.com/[EMAIL PROTECTED]
