SQL cells

SQL syntax highlighting.

Line numbering (can be enabled/disabled in the settings, see below). Clicking a line number selects the entire line.

Autocompletion — when you type an SQL query, a list of possible keywords, clauses, functions, database/table/column names is displayed based on the characters already entered. This feature can be enabled/disabled by the corresponding setting; and even if it is turned off, you can open suggestions by pressing Ctrl+Space.

Automatic closing of brackets and quotation marks as you type (controlled by the corresponding setting).

Automatic increasing/decreasing the editor based on the size of the SQL code. The editor can also be resized manually by dragging its bottom border. Double-clicking the resize handle at the bottom of the editor returns it to the size determined automatically depending on the size of the query text.

use consistent indents and spaces to structure the code logically;

place major clauses (such as SELECT, FROM, WHERE, JOIN, etc.) on separate lines;

use uppercase for all SQL keywords (for example, SELECT, AS, FROM, GROUP BY, ORDER BY);

break long lists of columns to be selected into lines — one column per line.

Font Size — font size (the setting does not affect text cells);

Word Wrap — enables/disables line wrapping (in text cells, line wrapping is always enabled, regardless of this setting);

Line Numbers — enables/disables line numbering (in text cells, row numbering can only be enabled for editing mode);

Auto-Close Brackets & Quotes — enables/disables automatic closing of brackets and quotation marks as you type;

SQL Autocomplete — enables/disables automatic display of suggestions for completing keywords, functions, database/table/column names as you type SQL code.

Drag to reorder — drag and drop to move a cell relative to others within a notebook.

[n] — monotonically increasing cell creation identifier that does not change when the cell is moved or deleted.

Query settings — display settings that allow specifying limits to prevent the execution of a potentially resource-intensive query. Available settings:

Beautify SQL — apply automatic formatting to the query text.

AI — convert a text question into a SELECT SQL query (available if the Text-to-SQL service is installed in an ADQM cluster).

Query execution status:

n statements — number of queries in a cell (shown if a cell contains more than one query separated by semicolons).

n — history of query runs in a cell (shown if more than one run was performed; n — number of runs in the history).

Run — start query execution.

Toggle query — collapse/expand the query code editor (if the editor is collapsed, the beginning of the query text is displayed within the control panel).

Copy cell as image — copy a cell as an image.

Fullscreen — expand an SQL cell to full screen (or collapse it).

Duplicate cell — copy a cell.

Move up — move a cell one position above the previous one.

Move down — move a cell one position below the next one.

Add cell below — create a new cell of the same type below.

Delete cell — delete a cell.

Click Run in the control panel of the corresponding SQL cell. While a query is running, the Run button visually changes to Stop — clicking it cancels the query execution (KILL QUERY is sent).

Click an SQL cell to select it (a selected cell is highlighted in green) and press Ctrl+Enter (or Cmd+Enter).

Displaying column names and data types in headers.

Sorting data by clicking any column header.

Data search — filtering values in the result set by any column in real time using the Search results…​ field.

Smart formatting:

TYPE — chart type ( Line, Area, or Bar);

X-AXIS — a column whose values will be used for labels on the X axis;

Y-AXIS — one or more columns whose values will be used for labels on the Y axis.

Estimated — predicted query execution plan (similar to what EXPLAIN provides without actually running the query).

Actual — actual query execution plan (requires the system.processors_profile_log table). When this option is selected, general information about executing a query that read data from a table is displayed at the top of the panel:

Pipe — query execution pipeline the same as EXPLAIN PIPELINE outputs it.

Logical — logical query execution plan presented as a schema that includes only the steps, which relate to obtaining the requested data (reading/filtering/aggregating/sorting data). Uninformative blocks (for internal processes that ClickHouse runs "under the hood" to handle a query) are not shown here. Read the query plan from the bottom up.

The width of arrows between the blocks corresponds to the number of data rows obtained as a result of the previous step and received as input for the next step.

A block corresponding to the longest stage of query execution is highlighted in color.

Inside a query plan block or in its tooltip (shown when you hover over the block), you can view information about the number of data rows processed at this stage, the stage execution time, and how much of this time is relative to the total query execution time. A Read block displays the number of data parts and granules read from the table.

The line inside a block shows what time was spent on at this stage of the plan — in the tooltip, you can see which processes correspond to the different colors of the line and how long each process took to complete.

description of the possible issue (click Show step to highlight the plan step it affects);

recommendation on how to work around the issue and improve query performance (click How to fix to view the recommendation).

Table has no sorting key

Filter can’t use the sorting key

Filter skips the leading key column

Aggregation/Sort spilled to disk

Duration — query execution time;

Read — total number of rows read from all tables and table functions involved in a query;

Scanned — total number of bytes read from all tables and table functions involved in a query;

Memory — memory used by a query;

Threads — peak number of threads used simultaneously during query execution;

Node Breakdown — information about query execution on each host (when a query is executed in an ADQM cluster);

Details — detailed information including: query kind (SELECT, INSERT, etc.), number of rows written (for write queries), bytes written, number of rows in query results, size of query results in bytes, and names of databases/tables/columns accessed by the query.