The documentation of HIVE notes that LIMIT clause returns rows chosen at random. I have been running a SELECT table on a table with more than 800,000 records with LIMIT 1, but it always return me the same record.
I'm using the Shark distribution, and I am wondering whether this has got anything to do with this not expected behavior? Any thoughts would be appreciated.
Thanks,
Visakh
Even though the documentation states it returns rows at random, it's not actually true.
It returns "chosen rows at random" as it appears in the database without any where/order by clause. This means that it's not really random (or randomly chosen) as you would think, just that the order the rows are returned in can't be determined.
As soon as you slap a order by x DESC limit 5 on there, it returns the last 5 rows of whatever you're selecting from.
To get rows returned at random, you would need to use something like: order by rand() LIMIT 1
However it can have a speed impact if your indexes aren't setup properly. Usually I do a min/max to get the ID's on the table, and then do a random number between them, then select those records (in your case, would be just 1 record), which tends to be faster than having the database do the work, especially on a large dataset
To be safe you want to use
select * from table
distribute by rand()
sort by rand()
limit 10000;
The documentation may have been updated since this question was originally posted in 2014, but as of December,2017, the documentation now reads, "The following query returns 5 arbitrary customers".
In this case, "arbitrary" means method of selecting either is not deterministic or may not be worth the trouble to document. In other words, you shouldn't count on it as a reliable method for getting specific subset of records (e.g., for sampling). You should only use the Limit clause without an Order By clause if you are looking for expediency and want to get a small result set as quickly as possible (e.g., for QA purposes). Otherwise, use one of Order By, Cluster By, or Distribute By/Sort By as appropriate.
Related
I am using T-SQL and in the book T-SQL Fundamentals of Itzik Ben Gan, he said that Select clause is processed logically before the Order by clause.
I agree on this, but I want to know how the select is processed before the ORDER BY, when the TOP is in the select and it needs the result of the order by first?
Without rewriting a lot of what's already been written before:
https://learn.microsoft.com/en-us/sql/t-sql/queries/select-transact-sql
Short version: Imagine that SQL Server creates virtual tables during query execution. Those tables and their values are passed, step-by-step, through a logical process that determines your end result. The goal is to "fetch" the minimum number of rows from the beginning, and thereby filter out as few rows as possible. After all, why "fetch" 100,000 rows if you only want to see 2?
In the case of a TOP clause, you're only going to see those TOP x rows, but that doesn't mean that they are the only rows that SQL Server checked during the query execution.
On the contrary - if you're looking for the TOP x rows by some column value, then clearly SQL Server needs to make sure that it first analyzes the values for that column, orders them accordingly, and can only then present you with the TOP x rows. This is why having proper indexes can make such a difference when executing these sorts of queries.
This is very different from the WHERE clause, which can happen earlier on, because a value either = X, or it doesn't; so when scanning a table, SQL Server can know for every single row whether or not that row should be included in the final result set. With a TOP clause, this is not necessarily the case. Halfway through the query process, it doesn't necessarily know if there are more unread rows that should be included in that TOP or not - it can only know once the rows have been selected in accordance with all previous conditions, then ordered by your ORDER BY clause. Finally, it knows which rows should be in those TOP x rows you asked for.
Notice that in Microsoft's documentation, they explicitly state that the logical processing order can change from query to query, especially in extenuating circumstances (i.e., a VIEW that uses CONVERT(), or depending upon the indexes for a given table.)
Suppose, if following rows are inserted in chronological order into a table:
row1, row2, row3, row4, ..., row1000, row1001.
After a while, we delete/remove the latest row1001.
As in this post: How to get Top 5 records in SqLite?
If the below command is run:
SELECT * FROM <table> LIMIT 1;
Will it assuredly provide the "row1000"?
If no, then is there any efficient way to get the latest row(s)
without traversing through all the rows? -- i.e. without using
combination of ORDER BY and DESC.
[Note: For now I am using "SQLite", but it will be interesting for me to know about SQL in general as well.]
You're misunderstanding how SQL works. You're thinking row-by-row which is wrong. SQL does not "traverse rows" as per your concern; it operates on data as "sets".
Others have pointed out that relational database cannot be assumed to have any particular ordering, so you must use ORDER BY to explicitly specify ordering.
However (not mentioned yet is that), in order to ensure it performs efficiently, you need to create an appropriate index.
Whether you have an index or not, the correct query is:
SELECT <cols>
FROM <table>
ORDER BY <sort-cols> [DESC] LIMIT <no-rows>
Note that if you don't have an index the database will load all data and probably sort in memory to find the TOP n.
If you do have the appropriate index, the database will use the best index available to retrieve the TOP n rows as efficiently as possible.
Note that the sqllite documentation is very clear on the matter. The section on ORDER BY explains that ordering is undefined. And nothing in the section on LIMIT contradicts this (it simply constrains the number of rows returned).
If a SELECT statement that returns more than one row does not have an ORDER BY clause, the order in which the rows are returned is undefined.
This behaviour is also consistent with the ANSI standard and all major SQL implementations. Note that any database vendor that guaranteed any kind of ordering would have to sacrifice performance to the detriment of queries trying to retrieve data but not caring about order. (Not good for business.)
As a side note, flawed assumptions about ordering is an easy mistake to make (similar to flawed assumptions about uninitialised local variables).
RDBMS implementations are very likely to make ordering appear consistent. They follow a certain algorithm for adding data, a certain algorithm for retrieving data. And as a result, their operations are highly repeatable (it's what we love (and hate) about computers). So things repeatably look the same.
Theoretical examples:
Inserting a row results in the row being added to the next available free space. So data appears sequential. But an update would have to move the row to a new location if it no longer fits.
The DB engine might retrieve data sequentially from clustered index pages and seem to use clustered index as the 'natural ordering' ... until one day a page-split puts one of the pages in a different location. * Or a new version of the DMBS might cache certain data for performance, and suddenly order changes.
Real-world example:
The MS SQL Server 6.5 implementation of GROUP BY had the side-effect of also sorting by the group-by columns. When MS (in version 7 or 2000) implemented some performance improvements, GROUP BY would by default, return data in a hashed order. Many people blamed MS for breaking their queries when in fact they had made false assumptions and failed to ORDER BY their results as needed.
This is why the only guarantee of a specific ordering is to use the ORDER BY clause.
No. Table records have no inherent order. So it is undefined which row(s) to get with a LIMIT clause without an ORDER BY.
SQLite in its current implemantation may return the latest inserted row, but even if that were the case you must not rely on it.
Give a table a datetime column or some sortkey, if record order is important for you.
In SQL, data is stored in tables unordered. What comes out first one day might not be the same the next.
ORDER BY, or some other specific selection criteria is required to guarantee the correct value.
Maybe someone can explain this to me, but when querying a data table from Oracle, where multiple records exist for a key (say a customer ID), the record that appears first for that customer can vary if there is no implicit "order by" statement enforcing the order by say an alternate field such as a transaction type. So running the same query on the same table could yield a different record ordering than from 10 minutes ago.
E.g., one run could yield:
Cust_ID, Transaction_Type
123 A
123 B
Unless an "order by Transaction_Type" clause is used, Oracle could arbitrarily return the following result the next time the query is run:
Cust_ID, Transaction_Type
123 B
123 A
I guess I was under the impression that there was a database default ordering of rows in Oracle which (perhaps) reflected the physical ordering on the disk medium. In other words, an arbitrary order that is immutable and would guarantee the same result when a query is rerun.
Does this have to do with the optimizer and how it decides where to most efficiently retrieve the data?
Of course the best practice from a programming perspective is to force whatever ordering is required, I was just a little unsettled by this behavior.
The order of rows returned to the application from a SELECT statement is COMPLETELY ARBITRARY unless otherwise specified. If you want, need, or expect rows to return in a certain order, it is the user's responsibility to specify such an order.
(Caveat: Some versions of Oracle would implicitly sort data in ascending order if certain operations were used, such as DISTINCT, UNION, MINUS, INTERSECT, or GROUP BY. However, as Oracle has implemented hash sorting, the nature of the sort of the data can vary, and lots of SQL relying on that feature broke.)
There is no default ordering, ever. If you don't specify ORDER BY, you can get the same result the first 10000 times, then it can change.
Note that this is also true even with ORDER BY for equal values. For example:
Col1 Col2
1 1
2 1
3 2
4 2
If you use ORDER BY Col2, you still don't know if row 1 or 2 will come first.
Just image the rows in a table like balls in a basket. Do the balls have an order?
I dont't think there is any DBMS that guarantees an order if ORDER BY is not specified.
Some might always return the rows in the order they were inserted, but that is an implementation side effect.
Some execution plans might cause the result set to be ordered even without an ORDER BY, but again this is an implementation side-effect that you should not rely on.
If an ORDER BY clause is not present the database (not just Oracle - any relational database) is free to return rows in whatever order it happens to find them. This will vary depending on the query plan chosen by the optimizer.
If the order in which the rows are returned matters you must use an ORDER BY clause. You may sometimes get lucky and the rows will come back in the order you want them to be even without an ORDER BY, but there is no guarantee that A) you will get lucky on other queries, and B) the order in which the rows are returned tomorrow will be the same as the order in which they're returned today.
In addition, updates to the database product may change the behavior of queries. We had to scramble a bit when doing a major version upgrade last year when we found that Oracle 10 returned GROUP BY results in a different order than did Oracle 9. Reason - no ORDER BY clause.
ORDER BY - when the order of the returned data really matters.
The simple answer is that the SQL standard says that there is no default order for queries that do not have an ORDER BY statement, so you should never assume one.
The real reason would probably relate to the hashes assigned to each row as it is pulled into the record set. There is no reason to assume consistent hashing.
if you don't use ORDER BY, the order is arbitrary; however, dependent on phisical storage and memory aspects.
so, if you repeat the same query hundreds of times in 10 minutes, you will get almost the same order everytime, because probably nothing changes.
Things that could change the "noorder order" are:
the executing plan - if is changed(you have pointed
that)
inserts and deletes on the tables involved in the query.
other things like presence in memory of the rows.(other querys on other tables could influence that)
When you get into parallel data retrieval I/O isn't it possible to get different sequences on different runs, even with no change to the stored data?
That is, in a multiprocessing environment the order of completion of parallel threads is undefined and can vary with what else is happening on the same shared processor.
As I'm new to Oracle database engine, I noticed this behavior in my SELECT statements that has no ORDER BY.
I've been using Microsoft SQL Server for years now. SQL Server Engine always will retrieve data ordered by the table's "Clustered Index" which is basically the Primary Key Index. SQL Server will always insert new data in a sequential order based on the clustered index.
So when you perform a select on a table without order by in SQL Server, it will always retrieve data ordered by primary key value.
ORDER BY can cause serious performance overhead, that's why you do not want to use it unless you are not happy with inconsistent results order.
I ended up with a conclusion that in ALL my Oracle queries I must use ORDER BY or I will end up with unpredicted order which will greatly effect my end-user reports.
I was wondering if its faster to retrieve the "count" of the number of rows or to retrieve just 1 row using limit. The purpose being to see whether theres any row when given certain Where conditions.
A count is always an expensive query because it will take a full table scan. You requirements are not really clear to me, but if you just want to see whether there is any data it would be cheaper to do a regular select with a limit to 1.
A count must physically count all rows that match your criteria, which is unnecessary work as you don't care about the number.
Look at using EXISTS.
It think it depends on which storage engine you use for the database...
Anywawy, the good practise to test wether there are or not results is to check the return value from the feth() function!
[I'm using Oracle as an example here, but the same concepts apply usually across the board.]
COUNT has to physically identify all the rows that will be returned. Depending on the complexity of the query, the query plan could require a table scan on one or more of the tables of the query. You'd need to do an EXPLAIN PLAN to know for sure.
Returning a single row may require the same processing if an ORDER BY is required. The database can't just give you the first row until
all the rows are identified and
the rows have been sorted.
Also, depending on the number of rows being returned, the complexity of the ORDER BY, and the SGA size, a temporary table might need to be created (which causes all sorts of other overhead).
If there's no ORDER BY, a single row should be faster because as soon as the data identifies a single row to return, it's done. That said, you're not guaranteed from one execution to the next that the rows are returned in the same order, so usually an ORDER BY is involved.
It's faster to retrieve the count of the number of rows
Kind of a whimsical question, always something I've wondered about and I figure knowing why it does what it does might deepen my understanding a bit.
Let's say I do "SELECT TOP 10 * FROM TableName". In short timeframes, the same 10 rows come back, so it doesn't seem random. They weren't the first or last created. In my massive sample size of...one table, it isn't returning the min or max auto-incrementing primary key value.
I also figure the problem gets more complex when taking joins into account.
My database of choice is MSSQL, but I figure this might be an interesting question regardless of the platform.
If you do not supply an ORDER BY clause on a SELECT statement you will get rows back in arbitrary order.
The actual order is undefined, and depends on which blocks/records are already cached in memory, what order I/O is performed in, when threads in the database server are scheduled to run, and so on.
There's no rhyme or reason to the order and you should never base any expectations on what order rows will be in unless you supply an ORDER BY.
If they're not ordered by the calling query, I believe they're just returned in the order they were read off disk. This may vary because of the types of joins used or the indexes that looked up the values.
You can see this if the table has a clustered index on it (and you're just selecting - a JOIN can re-order things) - a SELECT will return the rows in clustered-index-order, even without an ORDER BY clause.
There is a very detailed explanation with examples here: http://sqlserverpedia.com/blog/sql-server-bloggers/its-the-natural-order-of-things-not/
"How do database servers decide which order to return rows without any “order by” statements?"
They simply do not take any "decision" with respect to ordering. They see the user doesn't care about ordering, and so they don't care either. And thus they simply go out to find the requested rows. The order in which they find them is normally the order in which you get them. That order depends on user-unpredictable things like the chosen physical access paths, ordering of physical records inside the database's physical files, etc. etc.
Don't let yourself be misled by the ordering as you get it, in the case where you didn't explicitly specify an ordering in your query. If you don't specify an ordering in your query, no ordering in the result set is guaranteed, even if in practice results seem to suggest that some ordering appears to be adhered to by the server.