You can query data in Druid datasources using Druid SQL. Druid translates SQL queries into its [native query language](querying.md). To learn about translation and how to get the best performance from Druid SQL, see [SQL query translation](sql-translation.md).
For information on permissions needed to make SQL queries, see [Defining SQL permissions](../operations/security-user-auth.md#sql-permissions).
This topic introduces Druid SQL syntax.
For more information and SQL querying options see:
- [Data types](./sql-data-types.md) for a list of supported data types for Druid columns.
- [Aggregation functions](./sql-aggregations.md) for a list of aggregation functions available for Druid SQL SELECT statements.
- [Scalar functions](./sql-scalar.md) for Druid SQL scalar functions including numeric and string functions, IP address functions, Sketch functions, and more.
- [SQL multi-value string functions](./sql-multivalue-string-functions.md) for operations you can perform on string dimensions containing multiple values.
- [Query translation](./sql-translation.md) for information about how Druid translates SQL queries to native queries before running them.
The UNNEST clause unnests array values. It's the SQL equivalent to the [unnest datasource](./datasource.md#unnest). The source for UNNEST can be an array or an input that's been transformed into an array, such as with helper functions like MV_TO_ARRAY or ARRAY.
The following is the general syntax for UNNEST, specifically a query that returns the column that gets unnested:
```sql
SELECT column_alias_name FROM datasource, UNNEST(source_expression1) AS table_alias_name1(column_alias_name1), UNNEST(source_expression2) AS table_alias_name2(column_alias_name2), ...
```
* The `datasource` for UNNEST can be any Druid datasource, such as the following:
* A table, such as `FROM a_table`.
* A subset of a table based on a query, a filter, or a JOIN. For example, `FROM (SELECT columnA,columnB,columnC from a_table)`.
* The `source_expression` for the UNNEST function must be an array and can come from any expression. If the dimension you are unnesting is a multi-value dimension, you have to specify `MV_TO_ARRAY(dimension)` to convert it to an implicit ARRAY type. You can also specify any expression that has an SQL array datatype. For example, you can call UNNEST on the following:
*`ARRAY[dim1,dim2]` if you want to make an array out of two dimensions.
*`ARRAY_CONCAT(dim1,dim2)` if you want to concatenate two multi-value dimensions.
* The `AS table_alias_name(column_alias_name)` clause is not required but is highly recommended. Use it to specify the output, which can be an existing column or a new one. Replace `table_alias_name` and `column_alias_name` with a table and column name you want to alias the unnested results to. If you don't provide this, Druid uses a nondescriptive name, such as `EXPR$0`.
Keep the following things in mind when writing your query:
- You must include the context parameter `"enableUnnest": true`.
- You can unnest multiple source expressions in a single query.
- Notice the comma between the datasource and the UNNEST function. This is needed in most cases of the UNNEST function. Specifically, it is not needed when you're unnesting an inline array since the array itself is the datasource.
- If you view the native explanation of a SQL UNNEST, you'll notice that Druid uses `j0.unnest` as a virtual column to perform the unnest. An underscore is added for each unnest, so you may notice virtual columns named `_j0.unnest` or `__j0.unnest`.
- UNNEST preserves the ordering of the source array that is being unnested.
For examples, see the [Unnest arrays tutorial](../tutorials/tutorial-unnest-arrays.md).
The UNNEST function has the following limitations:
- The function does not remove any duplicates or nulls in an array. Nulls will be treated as any other value in an array. If there are multiple nulls within the array, a record corresponding to each of the nulls gets created.
- Arrays inside complex JSON types are not supported.
- You cannot perform an UNNEST at ingestion time, including SQL-based ingestion using the MSQ task engine.
The WHERE clause refers to columns in the FROM table, and will be translated to [native filters](filters.md). The
WHERE clause can also reference a subquery, like `WHERE col1 IN (SELECT foo FROM ...)`. Queries like this are executed
as a join on the subquery, described in the [Query translation](sql-translation.md#subqueries) section.
Strings and numbers can be compared in the WHERE clause of a SQL query through implicit type conversion.
For example, you can evaluate `WHERE stringDim = 1` for a string-typed dimension named `stringDim`.
However, for optimal performance, you should explicitly cast the reference number as a string when comparing against a string dimension:
```
WHERE stringDim = '1'
```
Similarly, if you compare a string-typed dimension with reference to an array of numbers, cast the numbers to strings:
```
WHERE stringDim IN ('1', '2', '3')
```
Note that explicit type casting does not lead to significant performance improvement when comparing strings and numbers involving numeric dimensions since numeric dimensions are not indexed.
## GROUP BY
The GROUP BY clause refers to columns in the FROM table. Using GROUP BY, DISTINCT, or any aggregation functions will
trigger an aggregation query using one of Druid's [three native aggregation query types](sql-translation.md#query-types). GROUP BY
can refer to an expression or a select clause ordinal position (like `GROUP BY 2` to group by the second selected
column).
The GROUP BY clause can also refer to multiple grouping sets in three ways. The most flexible is GROUP BY GROUPING SETS,
for example `GROUP BY GROUPING SETS ( (country, city), () )`. This example is equivalent to a `GROUP BY country, city`
followed by `GROUP BY ()` (a grand total). With GROUPING SETS, the underlying data is only scanned one time, leading to
better efficiency. Second, GROUP BY ROLLUP computes a grouping set for each level of the grouping expressions. For
example `GROUP BY ROLLUP (country, city)` is equivalent to `GROUP BY GROUPING SETS ( (country, city), (country), () )`
and will produce grouped rows for each country / city pair, along with subtotals for each country, along with a grand
total. Finally, GROUP BY CUBE computes a grouping set for each combination of grouping expressions. For example,
`GROUP BY CUBE (country, city)` is equivalent to `GROUP BY GROUPING SETS ( (country, city), (country), (city), () )`.
Grouping columns that do not apply to a particular row will contain `NULL`. For example, when computing
`GROUP BY GROUPING SETS ( (country, city), () )`, the grand total row corresponding to `()` will have `NULL` for the
"country" and "city" columns. Column may also be `NULL` if it was `NULL` in the data itself. To differentiate such rows,
you can use `GROUPING` aggregation.
When using GROUP BY GROUPING SETS, GROUP BY ROLLUP, or GROUP BY CUBE, be aware that results may not be generated in the
order that you specify your grouping sets in the query. If you need results to be generated in a particular order, use
the ORDER BY clause.
## HAVING
The HAVING clause refers to columns that are present after execution of GROUP BY. It can be used to filter on either
grouping expressions or aggregated values. It can only be used together with GROUP BY.
## ORDER BY
The ORDER BY clause refers to columns that are present after execution of GROUP BY. It can be used to order the results
based on either grouping expressions or aggregated values. ORDER BY can refer to an expression or a select clause
ordinal position (like `ORDER BY 2` to order by the second selected column). For non-aggregation queries, ORDER BY
can only order by the `__time` column. For aggregation queries, ORDER BY can order by any column.
## LIMIT
The LIMIT clause limits the number of rows returned. In some situations Druid will push down this limit to data servers,
which boosts performance. Limits are always pushed down for queries that run with the native Scan or TopN query types.
With the native GroupBy query type, it is pushed down when ordering on a column that you are grouping by. If you notice
that adding a limit doesn't change performance very much, then it's possible that Druid wasn't able to push down the
limit for your query.
## OFFSET
The OFFSET clause skips a certain number of rows when returning results.
If both LIMIT and OFFSET are provided, then OFFSET will be applied first, followed by LIMIT. For example, using
LIMIT 100 OFFSET 10 will return 100 rows, starting from row number 10.
Together, LIMIT and OFFSET can be used to implement pagination. However, note that if the underlying datasource is
modified between page fetches, then the different pages will not necessarily align with each other.
There are two important factors that can affect the performance of queries that use OFFSET:
- Skipped rows still need to be generated internally and then discarded, meaning that raising offsets to high values
can cause queries to use additional resources.
- OFFSET is only supported by the Scan and GroupBy [native query types](sql-translation.md#query-types). Therefore, a query with OFFSET
will use one of those two types, even if it might otherwise have run as a Timeseries or TopN. Switching query engines
The UNION ALL operator fuses multiple queries together. Druid SQL supports the UNION ALL operator in two situations: top-level and table-level, as described below. Queries that use UNION ALL in any other way will fail.
In top-level queries, you can use UNION ALL at the very top outer layer of the query - not in a subquery, and not in the FROM clause. The underlying queries run sequentially. Druid concatenates their results so that they appear one after the other.
In table-level queries, you must use UNION ALL in a subquery in the FROM clause, and create the lower-level subqueries that are inputs to the UNION ALL operator as simple table SELECTs. You can't use features like expressions, column aliasing, JOIN, GROUP BY, or ORDER BY in table-level queries.
The query runs natively using a [union datasource](datasource.md#union).
At table-level queries, you must select the same columns from each table in the same order, and those columns must either have the same types, or types that can be implicitly cast to each other (such as different numeric types). For this reason, it is generally more robust to write your queries to select specific columns. If you use `SELECT *`, you must modify your queries if a new column is added to one table but not to the others.
With table-level UNION ALL, the rows from the unioned tables are not guaranteed to process in any particular order. They may process in an interleaved fashion. If you need a particular result ordering, use [ORDER BY](#order-by) on the outer query.
For the legacy plan, be careful when interpreting EXPLAIN PLAN output, and use [request logging](../configuration/index.md#request-logging) if in doubt.
Request logs show the exact native query that will be run. Alternatively, to see the native query plan, set `useNativeQueryExplain` to true in the query context.
