Re: [PHP-DB] Slooooow query in MySQL.
In addition to Chris's suggestions you should also alter the homeid column (set default to NULL and update the whole database which shouldn't be a problem) so you don't have to do a double check on the same column. I would also suggest that the TSCN_MEDIAid column should be an int not a varchar. Aleksander Chris wrote: Rob Adams wrote: select h.addr, h.city, h.county, h.state, h.zip, 'yes' as show_prop, h.askingprice, '' as year_built, h.rooms, h.baths, '' as apt, '' as lot, h.sqft, h.listdate, '' as date_sold, h.comments, h.mlsnum, r.agency, concat(r.fname, ' ', r.lname) as rname, r.phone as rphone, '' as remail, '' as status, '' as prop_type, ts.TSCNfile as picture, h.homeid as homeid, 'yes' as has_virt from ProductStatus ps, home h, realtor r, ProductBin pb left join TourScene ts on ts.TSCNtourId = pb.PBINid and ts.TSCN_MEDIAid = '3' where ps.PSTSstatus = 'posted' and pb.PBINid = PSTS_POid and h.id = pb.PBINid and h.listdate DATE_SUB(NOW(), INTERVAL 2 YEAR) and (h.homeid is not null and h.homeid '') and r.realtorid = pb.PBIN_HALOid limit {l1}, {l2} Here is the query. I didn't know that it needed to have an ORDER clause in it for the limit to work properly. I'll probably order by h.listdate If you don't have an ORDER BY clause then you're going to get inconsistent results. The database will never guarantee returning results in a set order unless you tell it to by specifying an order by clause. To speed up your query, make sure you have indexes on: TourScene(TSCNtourId, TSCN_MEDIAid) ProductBin(PBINid, PBIN_HALOid) home(id, listdate) realtor(realtorid) If you can't get it fast, then post the EXPLAIN output. -- PHP Database Mailing List (http://www.php.net/) To unsubscribe, visit: http://www.php.net/unsub.php
Re: [PHP-DB] Slooooow query in MySQL.
Stut wrote: Chris wrote: Stut wrote: Chris wrote: Rob Adams wrote: I have a query that I run using mysql that returns about 60,000 plus rows. It's been so large that I've just been testing it with a limit 0, 1 (ten thousand) on the query. That used to take about 10 minutes to run, including processing time in PHP which spits out xml from the query. I decided to chunk the query down into 1,000 row increments, and tried that. The script processed 10,000 rows in 23 seconds! I was amazed! But unfortunately it takes quite a bit longer than 6*23 to process the 60,000 rows that way (1,000 at a time). It takes almost 8 minutes. I can't figure out why it takes so long, or how to make it faster. The data for 60,000 rows is about 120mb, so I would prefer not to use a temporary table. Any other suggestions? This is probably more a db issue than a php issue, but I thought I'd try here first. Sounds like missing indexes or something. Use explain: http://dev.mysql.com/doc/refman/4.1/en/explain.html If that were the case I wouldn't expect limiting the number of rows returned to make a difference since the actual query is the same. Actually it can. I don't think mysql does this but postgresql does take the limit/offset clauses into account when generating a plan. http://www.postgresql.org/docs/current/static/sql-select.html#SQL-LIMIT Not really relevant to the problem though :P How many queries do you run with an order? But you're right, if there is no order by clause adding a limit probably will make a difference, but there must be an order by when you use limit to ensure the SQL engine doesn't give you the same rows in response to more than one of the queries. Oops, that was meant to say How many queries do you run *without* an order? -Stut -- http://stut.net/ -- PHP Database Mailing List (http://www.php.net/) To unsubscribe, visit: http://www.php.net/unsub.php
Re: [PHP-DB] Slooooow query in MySQL.
select h.addr, h.city, h.county, h.state, h.zip, 'yes' as show_prop, h.askingprice, '' as year_built, h.rooms, h.baths, '' as apt, '' as lot, h.sqft, h.listdate, '' as date_sold, h.comments, h.mlsnum, r.agency, concat(r.fname, ' ', r.lname) as rname, r.phone as rphone, '' as remail, '' as status, '' as prop_type, ts.TSCNfile as picture, h.homeid as homeid, 'yes' as has_virt from ProductStatus ps, home h, realtor r, ProductBin pb left join TourScene ts on ts.TSCNtourId = pb.PBINid and ts.TSCN_MEDIAid = '3' where ps.PSTSstatus = 'posted' and pb.PBINid = PSTS_POid and h.id = pb.PBINid and h.listdate DATE_SUB(NOW(), INTERVAL 2 YEAR) and (h.homeid is not null and h.homeid '') and r.realtorid = pb.PBIN_HALOid limit {l1}, {l2} Here is the query. I didn't know that it needed to have an ORDER clause in it for the limit to work properly. I'll probably order by h.listdate -- Rob Stut [EMAIL PROTECTED] wrote in message news:[EMAIL PROTECTED] Chris wrote: Stut wrote: Chris wrote: Rob Adams wrote: I have a query that I run using mysql that returns about 60,000 plus rows. It's been so large that I've just been testing it with a limit 0, 1 (ten thousand) on the query. That used to take about 10 minutes to run, including processing time in PHP which spits out xml from the query. I decided to chunk the query down into 1,000 row increments, and tried that. The script processed 10,000 rows in 23 seconds! I was amazed! But unfortunately it takes quite a bit longer than 6*23 to process the 60,000 rows that way (1,000 at a time). It takes almost 8 minutes. I can't figure out why it takes so long, or how to make it faster. The data for 60,000 rows is about 120mb, so I would prefer not to use a temporary table. Any other suggestions? This is probably more a db issue than a php issue, but I thought I'd try here first. Sounds like missing indexes or something. Use explain: http://dev.mysql.com/doc/refman/4.1/en/explain.html If that were the case I wouldn't expect limiting the number of rows returned to make a difference since the actual query is the same. Actually it can. I don't think mysql does this but postgresql does take the limit/offset clauses into account when generating a plan. http://www.postgresql.org/docs/current/static/sql-select.html#SQL-LIMIT Not really relevant to the problem though :P How many queries do you run with an order? But you're right, if there is no order by clause adding a limit probably will make a difference, but there must be an order by when you use limit to ensure the SQL engine doesn't give you the same rows in response to more than one of the queries. -Stut -- http://stut.net/ -- PHP Database Mailing List (http://www.php.net/) To unsubscribe, visit: http://www.php.net/unsub.php
Re: [PHP-DB] Slooooow query in MySQL.
Stut wrote: Stut wrote: Chris wrote: Stut wrote: Chris wrote: Rob Adams wrote: I have a query that I run using mysql that returns about 60,000 plus rows. It's been so large that I've just been testing it with a limit 0, 1 (ten thousand) on the query. That used to take about 10 minutes to run, including processing time in PHP which spits out xml from the query. I decided to chunk the query down into 1,000 row increments, and tried that. The script processed 10,000 rows in 23 seconds! I was amazed! But unfortunately it takes quite a bit longer than 6*23 to process the 60,000 rows that way (1,000 at a time). It takes almost 8 minutes. I can't figure out why it takes so long, or how to make it faster. The data for 60,000 rows is about 120mb, so I would prefer not to use a temporary table. Any other suggestions? This is probably more a db issue than a php issue, but I thought I'd try here first. Sounds like missing indexes or something. Use explain: http://dev.mysql.com/doc/refman/4.1/en/explain.html If that were the case I wouldn't expect limiting the number of rows returned to make a difference since the actual query is the same. Actually it can. I don't think mysql does this but postgresql does take the limit/offset clauses into account when generating a plan. http://www.postgresql.org/docs/current/static/sql-select.html#SQL-LIMIT Not really relevant to the problem though :P How many queries do you run with an order? But you're right, if there is no order by clause adding a limit probably will make a difference, but there must be an order by when you use limit to ensure the SQL engine doesn't give you the same rows in response to more than one of the queries. Oops, that was meant to say How many queries do you run *without* an order? Almost never - but my point was actually this sentence: The query planner takes LIMIT into account when generating a query plan, so you are very likely to get different plans (yielding different row orders) depending on what you use for LIMIT and OFFSET. -- Postgresql php tutorials http://www.designmagick.com/ -- PHP Database Mailing List (http://www.php.net/) To unsubscribe, visit: http://www.php.net/unsub.php
Re: [PHP-DB] Slooooow query in MySQL.
Rob Adams wrote: select h.addr, h.city, h.county, h.state, h.zip, 'yes' as show_prop, h.askingprice, '' as year_built, h.rooms, h.baths, '' as apt, '' as lot, h.sqft, h.listdate, '' as date_sold, h.comments, h.mlsnum, r.agency, concat(r.fname, ' ', r.lname) as rname, r.phone as rphone, '' as remail, '' as status, '' as prop_type, ts.TSCNfile as picture, h.homeid as homeid, 'yes' as has_virt from ProductStatus ps, home h, realtor r, ProductBin pb left join TourScene ts on ts.TSCNtourId = pb.PBINid and ts.TSCN_MEDIAid = '3' where ps.PSTSstatus = 'posted' and pb.PBINid = PSTS_POid and h.id = pb.PBINid and h.listdate DATE_SUB(NOW(), INTERVAL 2 YEAR) and (h.homeid is not null and h.homeid '') and r.realtorid = pb.PBIN_HALOid limit {l1}, {l2} Here is the query. I didn't know that it needed to have an ORDER clause in it for the limit to work properly. I'll probably order by h.listdate If you don't have an ORDER BY clause then you're going to get inconsistent results. The database will never guarantee returning results in a set order unless you tell it to by specifying an order by clause. To speed up your query, make sure you have indexes on: TourScene(TSCNtourId, TSCN_MEDIAid) ProductBin(PBINid, PBIN_HALOid) home(id, listdate) realtor(realtorid) If you can't get it fast, then post the EXPLAIN output. -- Postgresql php tutorials http://www.designmagick.com/ -- PHP Database Mailing List (http://www.php.net/) To unsubscribe, visit: http://www.php.net/unsub.php
Re: [PHP-DB] Slooooow query in MySQL.
Stut wrote: Chris wrote: Rob Adams wrote: I have a query that I run using mysql that returns about 60,000 plus rows. It's been so large that I've just been testing it with a limit 0, 1 (ten thousand) on the query. That used to take about 10 minutes to run, including processing time in PHP which spits out xml from the query. I decided to chunk the query down into 1,000 row increments, and tried that. The script processed 10,000 rows in 23 seconds! I was amazed! But unfortunately it takes quite a bit longer than 6*23 to process the 60,000 rows that way (1,000 at a time). It takes almost 8 minutes. I can't figure out why it takes so long, or how to make it faster. The data for 60,000 rows is about 120mb, so I would prefer not to use a temporary table. Any other suggestions? This is probably more a db issue than a php issue, but I thought I'd try here first. Sounds like missing indexes or something. Use explain: http://dev.mysql.com/doc/refman/4.1/en/explain.html If that were the case I wouldn't expect limiting the number of rows returned to make a difference since the actual query is the same. Actually it can. I don't think mysql does this but postgresql does take the limit/offset clauses into account when generating a plan. http://www.postgresql.org/docs/current/static/sql-select.html#SQL-LIMIT Not really relevant to the problem though :P -- Postgresql php tutorials http://www.designmagick.com/ -- PHP Database Mailing List (http://www.php.net/) To unsubscribe, visit: http://www.php.net/unsub.php
Re: [PHP-DB] Slooooow query in MySQL.
Rob Adams napsal(a): I have a query that I run using mysql that returns about 60,000 plus rows. It's been so large that I've just been testing it with a limit 0, 1 (ten thousand) on the query. That used to take about 10 minutes to run, including processing time in PHP which spits out xml from the query. I decided to chunk the query down into 1,000 row increments, and tried that. The script processed 10,000 rows in 23 seconds! I was amazed! But unfortunately it takes quite a bit longer than 6*23 to process the 60,000 rows that way (1,000 at a time). It takes almost 8 minutes. I can't figure out why it takes so long, or how to make it faster. The data for 60,000 rows is about 120mb, so I would prefer not to use a temporary table. Any other suggestions? This is probably more a db issue than a php issue, but I thought I'd try here first. 60k rows is not that much, I have tables with 500k rows and queries are running smoothly. Anyway we cannot help you if you do not post: 1. show create table 2. result of explain query 3. the query itself OKi98 -- PHP Database Mailing List (http://www.php.net/) To unsubscribe, visit: http://www.php.net/unsub.php
Re: [PHP-DB] Slooooow query in MySQL.
60k records shouldn't be a problem. Show us the query you're making and the table structure. OKi98 wrote: Rob Adams napsal(a): I have a query that I run using mysql that returns about 60,000 plus rows. It's been so large that I've just been testing it with a limit 0, 1 (ten thousand) on the query. That used to take about 10 minutes to run, including processing time in PHP which spits out xml from the query. I decided to chunk the query down into 1,000 row increments, and tried that. The script processed 10,000 rows in 23 seconds! I was amazed! But unfortunately it takes quite a bit longer than 6*23 to process the 60,000 rows that way (1,000 at a time). It takes almost 8 minutes. I can't figure out why it takes so long, or how to make it faster. The data for 60,000 rows is about 120mb, so I would prefer not to use a temporary table. Any other suggestions? This is probably more a db issue than a php issue, but I thought I'd try here first. 60k rows is not that much, I have tables with 500k rows and queries are running smoothly. Anyway we cannot help you if you do not post: 1. show create table 2. result of explain query 3. the query itself OKi98 -- PHP Database Mailing List (http://www.php.net/) To unsubscribe, visit: http://www.php.net/unsub.php
Re: [PHP-DB] Slooooow query in MySQL.
Chris wrote: Rob Adams wrote: I have a query that I run using mysql that returns about 60,000 plus rows. It's been so large that I've just been testing it with a limit 0, 1 (ten thousand) on the query. That used to take about 10 minutes to run, including processing time in PHP which spits out xml from the query. I decided to chunk the query down into 1,000 row increments, and tried that. The script processed 10,000 rows in 23 seconds! I was amazed! But unfortunately it takes quite a bit longer than 6*23 to process the 60,000 rows that way (1,000 at a time). It takes almost 8 minutes. I can't figure out why it takes so long, or how to make it faster. The data for 60,000 rows is about 120mb, so I would prefer not to use a temporary table. Any other suggestions? This is probably more a db issue than a php issue, but I thought I'd try here first. Sounds like missing indexes or something. Use explain: http://dev.mysql.com/doc/refman/4.1/en/explain.html If that were the case I wouldn't expect limiting the number of rows returned to make a difference since the actual query is the same. Chances are it's purely a data transfer delay. Do a test with the same query but only grab one of the fields - something relative small like a integer field - and see if that's significantly quicker. I'm betting it will be. If that is the problem you need to be looking at making sure you're only getting the fields you need. You may also want to look into changing the cursor type you're using although I'm not sure if that's possible with MySQL nevermind how to do it. -Stut -- http://stut.net/ -- PHP Database Mailing List (http://www.php.net/) To unsubscribe, visit: http://www.php.net/unsub.php
[PHP-DB] Slooooow query in MySQL.
I have a query that I run using mysql that returns about 60,000 plus rows. It's been so large that I've just been testing it with a limit 0, 1 (ten thousand) on the query. That used to take about 10 minutes to run, including processing time in PHP which spits out xml from the query. I decided to chunk the query down into 1,000 row increments, and tried that. The script processed 10,000 rows in 23 seconds! I was amazed! But unfortunately it takes quite a bit longer than 6*23 to process the 60,000 rows that way (1,000 at a time). It takes almost 8 minutes. I can't figure out why it takes so long, or how to make it faster. The data for 60,000 rows is about 120mb, so I would prefer not to use a temporary table. Any other suggestions? This is probably more a db issue than a php issue, but I thought I'd try here first. -- PHP Database Mailing List (http://www.php.net/) To unsubscribe, visit: http://www.php.net/unsub.php
Re: [PHP-DB] Slooooow query in MySQL.
Seeing the query would help. Are you using sub-queries? I believe that those can make the time go up exponentially. -- Kevin Murphy Webmaster: Information and Marketing Services Western Nevada College www.wnc.edu 775-445-3326 P.S. Please note that my e-mail and website address have changed from wncc.edu to wnc.edu. On Jul 19, 2007, at 2:19 PM, Rob Adams wrote: I have a query that I run using mysql that returns about 60,000 plus rows. It's been so large that I've just been testing it with a limit 0, 1 (ten thousand) on the query. That used to take about 10 minutes to run, including processing time in PHP which spits out xml from the query. I decided to chunk the query down into 1,000 row increments, and tried that. The script processed 10,000 rows in 23 seconds! I was amazed! But unfortunately it takes quite a bit longer than 6*23 to process the 60,000 rows that way (1,000 at a time). It takes almost 8 minutes. I can't figure out why it takes so long, or how to make it faster. The data for 60,000 rows is about 120mb, so I would prefer not to use a temporary table. Any other suggestions? This is probably more a db issue than a php issue, but I thought I'd try here first. -- PHP Database Mailing List (http://www.php.net/) To unsubscribe, visit: http://www.php.net/unsub.php
Re: [PHP-DB] Slooooow query in MySQL.
Rob Adams wrote: I have a query that I run using mysql that returns about 60,000 plus rows. It's been so large that I've just been testing it with a limit 0, 1 (ten thousand) on the query. That used to take about 10 minutes to run, including processing time in PHP which spits out xml from the query. I decided to chunk the query down into 1,000 row increments, and tried that. The script processed 10,000 rows in 23 seconds! I was amazed! But unfortunately it takes quite a bit longer than 6*23 to process the 60,000 rows that way (1,000 at a time). It takes almost 8 minutes. I can't figure out why it takes so long, or how to make it faster. The data for 60,000 rows is about 120mb, so I would prefer not to use a temporary table. Any other suggestions? This is probably more a db issue than a php issue, but I thought I'd try here first. Sounds like missing indexes or something. Use explain: http://dev.mysql.com/doc/refman/4.1/en/explain.html -- Postgresql php tutorials http://www.designmagick.com/ -- PHP Database Mailing List (http://www.php.net/) To unsubscribe, visit: http://www.php.net/unsub.php