EXPLAIN

    Description

    displays the query plan that the Greenplum planner generates for the supplied statement. Query plans are a tree plan of nodes. Each node in the plan represents a single operation, such as table scan, join, aggregation or a sort.

    Plans should be read from the bottom up as each node feeds rows into the node directly above it. The bottom nodes of a plan are usually table scan operations (sequential, index or bitmap index scans). If the query requires joins, aggregations, or sorts (or other operations on the raw rows) then there will be additional nodes above the scan nodes to perform these operations. The topmost plan nodes are usually the Greenplum Database motion nodes (redistribute, explicit redistribute, broadcast, or gather motions). These are the operations responsible for moving rows between the segment instances during query processing.

    The output of EXPLAIN has one line for each node in the plan tree, showing the basic node type plus the following cost estimates that the planner made for the execution of that plan node:

    • cost — measured in units of disk page fetches; that is, 1.0 equals one sequential disk page read. The first estimate is the start-up cost (cost of getting to the first row) and the second is the total cost (cost of getting all rows). Note that the total cost assumes that all rows will be retrieved, which may not always be the case (if using LIMIT for example).
    • rows — the total number of rows output by this plan node. This is usually less than the actual number of rows processed or scanned by the plan node, reflecting the estimated selectivity of any WHERE clause conditions. Ideally the top-level nodes estimate will approximate the number of rows actually returned, updated, or deleted by the query.

    It is important to note that the cost of an upper-level node includes the cost of all its child nodes. The topmost node of the plan has the estimated total execution cost for the plan. This is this number that the planner seeks to minimize. It is also important to realize that the cost only reflects things that the query optimizer cares about. In particular, the cost does not consider the time spent transmitting result rows to the client.

    EXPLAIN ANALYZE causes the statement to be actually executed, not only planned. The EXPLAIN ANALYZE plan shows the actual results along with the planner’s estimates. This is useful for seeing whether the planner’s estimates are close to reality. In addition to the information shown in the EXPLAIN plan, will show the following additional information:

    • The total elapsed time (in milliseconds) that it took to run the query.

    • The number of workers (segments) involved in a plan node operation. Only segments that return rows are counted.

    • The maximum number of rows returned by the segment that produced the most rows for an operation. If multiple segments produce an equal number of rows, the one with the longest time to end is the one chosen.

    • The time (in milliseconds) it took to retrieve the first row from the segment that produced the most rows, and the total time taken to retrieve all rows from that segment. The <time> to first row may be omitted if it is the same as the <time> to end.

    Important: Keep in mind that the statement is actually executed when EXPLAIN ANALYZE is used. Although EXPLAIN ANALYZE will discard any output that a SELECT would return, other side effects of the statement will happen as usual. If you wish to use EXPLAIN ANALYZE on a DML statement without letting the command affect your data, use this approach:

    name

    The name of the prepared statement to execute.

    parameter

    The actual value of a parameter to the prepared statement. This must be an expression yielding a value that is compatible with the data type of this parameter, as was determined when the prepared statement was created.

    Notes

    In order to allow the query optimizer to make reasonably informed decisions when optimizing queries, the ANALYZE statement should be run to record statistics about the distribution of data within the table. If you have not done this (or if the statistical distribution of the data in the table has changed significantly since the last time was run), the estimated costs are unlikely to conform to the real properties of the query, and consequently an inferior query plan may be chosen.

    For more information about query profiling, see “Query Profiling” in the Greenplum Database Administrator Guide. For more information about resource queues, see “Resource Management with Resource Queues” in the Greenplum Database Administrator Guide.

    To illustrate how to read an EXPLAIN query plan, consider the following example for a very simple query:

    If we read the plan from the bottom up, the query optimizer starts by doing a sequential scan of the names table. Notice that the WHERE clause is being applied as a filter condition. This means that the scan operation checks the condition for each row it scans, and outputs only the ones that pass the condition.

    The results of the scan operation are passed up to a gather motion operation. In Greenplum Database, a gather motion is when segments send rows up to the master. In this case we have 2 segment instances sending to 1 master instance (2:1). This operation is working on slice1 of the parallel query execution plan. In Greenplum Database a query plan is divided into slices so that portions of the query plan can be worked on in parallel by the segments.

    The estimated startup cost for this plan is 00.00 (no cost) and a total cost of disk page fetches. The planner is estimating that this query will return one row.

    Compatibility

    There is no EXPLAIN statement defined in the SQL standard.

    Parent topic: SQL Command Reference