oracle - Is an Index Organized Table appropriate here? -

-

I was recently reading about Oracle Index Organized Tables (IOT), but I'm not sure they used When I am quite understandable. So I have a small table: 

  create table categories (ID VARCHAR2 (36), group VARCHAR2 (100), category VARCHAR2 (100) create unique index (group, category, id) COMPRESS 2;

ID column is a foreign key entries from another table and my usual query is:

Select 
  e.id, e.time, from e.title entries, categories C where e.id = c.id and for example =? And c.category =? E.time  ORDER by code>  
 The table is properly indexed.  
 There are millions of rows in this table (currently 16 m) and the query is currently really smelling (note: I put it on a single page to query I have only wrapped 20 already, but I have been given up for simplicity).  
 Since I basically index entire table, does this table IOT 
  Edit :  
  Create table entries (ID VARCHAR2 (36), TIMESTEMP time by popular demand , Group VARCHAR2 (100), title VARCHAR2 (500), ....) Create index (group, time) sec 1;  
 
 My actual question I do not think depends on it though. In fact if you have a table with some columns (3 in this example) and you are planning to put a composite index on all three lines, then what is the reason for not using IOT? IOTs are great for many purposes, in this case, where you have all the columns (or most of them) in the column.

On any indicator - but the benefit is only if you do not additional index - the idea that the table is automatically an index, so put the column in that order so that you want to put the index inside. In your case, you are reaching the category based on the ID, so it makes sense for the first column. So effectively you got an index on (id, group, category). I do not know why you want an additional index on (group, category, id).

Your query: 

  SELECT e.id, e.time, entries from e.title e, categories C WHERE e.id = c.id and for example =? And c.category =? E. Order by Time

You are joining tables from IDs, but you do not have any indices on entries.id - so the query is probably joining a hash or sort merge

If you want to score points on a page (only interested in a small number of rows), then you want to get first the rows as soon as possible; To do this, you probably need a nested loop on entries, such as: 

  Entrance table to be entered by Roaded - Indication index range scan - (ENTRIES.group, Time Index) Access Table ROWID BY - CATEGORIES INDEX RANGE SCAN - (Index on CATEGORIES.ID)

Since ID is included in CATEGORIES, you need an index on ID; If you make it an IOT, and make the leading column ID, then it can be enough.

The performance of the plan shown above depends on how many rows match the given "group" - that is how an average "group" is selected.


Comments

Post a Comment

Popular posts from this blog

MySql variables and php -

-
I am getting an error with php in order to format mysql_query () in php to format this string What is the right way? SELECT SELECT SELECT (*) Agents @AgentCount; User_Agent_English, User_Angent_Oursal, US.___Gent_Aginal, COUNT (User_Agent_Percade) ACC, COUNT (*) / (AgentCount) Percentage by agents user_agent_preser order DEC limit limit 50; In PHP, how am i setting $ query $ query = "SELECT count (*) agents @ agentcount; SELECT user_agent_parsed, User_agent_original , COUNT (user_agent_parsed) ascount, COUNT (*) / (@AgentCount) as a percentage by user_agent_parced by URI date DESC limit 50 "; This exact query will work properly if I put it in MySQL directly through a command line session. Do I need to release two separate php calls from mysql_query () and store the results first? I get the following error: There is an error in your SQL syntax; Instead of using a sub-selection and instead of using the 'SELECT user_agent_parsed, user_agent_original, ne...
Read more

url rewriting - How to implement the returnurl like SO in PHP? -

-
This URL should work even after the rewrite. Is this possible? Otherwise, I can only hardcoded everything If you are trying to create a login system with redirection You need to use the HTTP_REFERER header to find out where the user came from so that you can send it back there. The best way to do this is to use something like this in your login form: // - Login php & Lt; Input type = "password" name = "password" /> & Lt; Input type = "hidden" name = "returnrelale" value = "& lt ;? php echo $ redirectURL ;? & gt;" / & Gt; & Lt; / Form & gt; Please note that some ISP, browser or proxy will delete the referrer header, so you can not rely on each user to work fully. If you use a statement, to check whether the header exists or not, you can send those users back to the homepage or whatever. Edit If you are looking to use $ _SERVER ['REQUEST_URI'], if you are running Apache It sho...
Read more

Which Python client library should I use for CouchdB? -

-
I am starting to experiment with CouchDB because it looks like we have the right solution to certain problems That all the work will be on a new project, which does not have any legacy dependency, which the client library will suggest you use, and why? If we overlap any of the OS. Freeband is available only in the collection of its ports, but asks to use the library's project website. None of those people are Ubuntu, which is only with ships, since those two OSAs are not common in libraries, I am probably installing something from source (and hopefully if I have time, then Ubuntu and FreeBSD People pack the package). For those interests, I want to use CouchDB as a convenient intermediate storage space for data passed between different services - think of a message bus system, but less formalities with. For example, we have daemons that download and pars the web page, then send an interesting bit to other daemons for further processing. Many of them are ill-defined until the...
Read more