Using the Management Portal SQL Interface

All SQL operations occur within a specific namespace. Therefore, you must first specify which namespace you wish to use by clicking the name of the current namespace displayed at the top of the SQL interface page. This displays the list of available namespaces, from which you can make your selection.

You can set your Management Portal default namespace. From the Management Portal select System Administration, Security, Users. Click the name of the desired user. This allows you to edit the user definition. From the General tab, select a Startup Namespace from the drop-down list. Click Save. If no startup namespace is selected, it defaults to %SYS.

Many of the Management Portal SQL operations are automatically customized for each user. If you set a filter, maximum, mode, or other option in the Execute Query tab or the SQL Statements tab, this user-specified value is retained for future use. When the same user activates the Management Portal, the user’s prior settings are shown. Restarting InterSystems IRIS returns all options to default values.

Namespace selection is not customized. It reverts to the user definition Startup Namespace.

For details on using Filter options, see Filtering Schema Contents.

The Execute Query text box allows you to write not only SELECT and CALL queries, but most SQL statements, including DDL statements such as CREATE TABLE, and DML statements such as INSERT, UPDATE, and DELETE.

You can specify SQL code in the Execute Query text box using the following:

• Type (or paste) the SQL code into the text box. The SQL code area does not colorize SQL text or provide any syntax or existence validation. However, it does provide automatic spelling verification. You can erase the contents of the text box using the X icon.

• Use the Show History list to select a prior SQL statement. The selected statement is copied into the text box. Upon execution, this statement moves to the top of the Show History list. Note that Show History lists all previously executed statements, including those that failed execution.

• Use Table Drag and Drop to construct SQL code in the text box.

• You can use the Query Builder, rather than the Execute Query text box, to specify and execute a SELECT query. A SELECT query executed using Query Builder is not shown in Execute Query or listed in Show History.

SQL code in the Execute Query text box can include:

• ? Input Parameters. If you specify input parameters, such as TOP ? or WHERE Age BETWEEN ? AND ?, the Execute button displays the Enter Parameter Value for Query window, with entry fields for each input parameter in the order specified in the query. For further details on ? input parameters, refer to Executing an SQL Statement.

• Whitespace Characters. You can specify multiple blank spaces, single and multiple line returns. The tab key is disabled; when copying code into the SQL code area, existing tabs are converted to single blank spaces. Line returns and multiple blank spaces are not retained.

• Comments. The SQL code area supports single-line and multiline comments. Comments are retained and shown in the Show History display. Comments are not shown in the Show Plan Statement Text display or in cached queries.

• Queries that Return Multiple Result Sets.

After writing SQL code in the text box, you can click the Show Plan button to check the SQL code without executing the SQL code. If the code is valid, Show Plan displays a Query Plan. If the code is invalid, Show Plan displays an SQLCODE error value and message. You can also use the Show Plan button to display this information for the most-recently-executed SQL code.

To execute the SQL code, click the Execute button.

The SQL execution interface has the following options:

• The Select Mode drop-down list with a SELECT specifies the format that the query should use to supply data values (for example, in the WHERE clause) and to display data values in the query result set. The options are Display Mode (the default), ODBC Mode, and Logical Mode. For further details on these options, refer to Data Display Options.

  The Select Mode drop-down list with an INSERT or UPDATE allows you to specify whether input data will be converted from display format to logical storage format. For this data conversion to occur, the SQL code must have been compiled with a select mode of RUNTIME. At execution time, the Select Mode drop-down list must be set to Logical Mode. For further details, refer to the INSERT or UPDATE statements.

  Select Mode is meaningful for data types whose Logical storage format differs from the desired display format (Display or ODBC), such as InterSystems IRIS dates and times and ObjectScript %List structured data.

• The Max field allows you to limit how many rows of data to return from a query. It can be set to any positive integer, including 0. Once you set Max, that value is used for all queries for the duration of the session, unless explicitly changed. The default is 1000. The maximum value is 100,000, which is the default if you enter no value (set Max to null), enter a value greater than 100,000, or a non-numeric value. You can also limit the number of rows of data to return by using a TOP clause. Max has no effect on other SQL statements, such as DELETE.

If you click the more option, the SQL execution interface displays the following additional options:

• Dialect: the dialect of SQL code. Available values are IRIS, Sybase, and MSSQL. The default is IRIS. Sybase and MSSQL are described in the InterSystems Transact-SQL (TSQL) Migration Guide. Note that the dialect you select becomes the user customized default the next time you access the Management Portal.

  If Dialect is Sybase or MSSQL, you can specify multiple SQL commands in the Execute Query text box. All insert, delete, and update commands are executed first, then SELECT commands in the order specified. Multiple result sets are returned in separate tabs.

• Row Number: a check box specifying whether to include a row count number for each row in the result set display. Row Number is a sequential integer assigned to each row in the result set. This is simply a numbering of the returned rows, it does not correspond to either the RowID or the %VID. The row number column header name is #. The default is to not display row numbers.

• Execute Query in the foreground: a check box specifying whether or not to run the query in the foreground. Simple queries run in the foreground are often significantly faster than those run in the background. However, long queries run in the foreground may cause the Management Portal to be unresponsive during query execution. The default is to run all queries in the background.

All of these options are user customized.

The Show Plan button displays the Statement Text and the Query Plan including the relative cost (overhead) of the current query plan for the query in the page’s text box. You can invoke Show Plan from either the Execute Query or Show History interface.

A query plan is generated when a query is Prepared (compiled); this occurs when you write a query and select the Show Plan button. You do not have to execute a query to show its query plan.

When viewing a query plan through the Management Portal, high level descriptions of how the plan processes modules and subqueries are hyperlinked to the sections that explain in detail how these parts of the query are processed, allowing you to easily traverse the plan.

Show Plan displays an SQLCODE and error message when invoked for an invalid query.

After writing SQL code in the Execute Query text box, you can execute the code by clicking the Execute button. This either successfully executes the SQL statement and displays the results below the code window, or the SQL code fails. If the SQL code fails, it displays an error message (in red) below the code window; pressing the Show Plan button displays the SQLCODE error and error message.

Execute Query SQL code execution is performed as a background process. While the code is executing, the Execute button is replaced by a Cancel button. This allows you to cancel execution of a long-running query.

Click Show History to list prior SQL statements executed during the current session. Show History lists all SQL statements invoked from this interface, both those successfully executed and those whose execution failed. By default, SQL statements are listed by Execution Time, with the most recently executed appearing at the top of the list. You can click on any of the column headings to order the SQL statements in ascending or descending order by column values. Executing an SQL Statement from the Show History listing updates its Execution Time (local date and time stamp), and increments its Count (number of times executed).

You can filter the Show History listing, as follows: in the Filter box specify a string then press the Tab key. Only those history items that contain that string will be included in the refreshed listing. The filter string can either be a string found in the SQL Statement column (such as a table name), or it can be a string found in the Execution Time column (such as a date). The filter string is not case-sensitive. A filter string remains in effect until you explicitly change it.

You can modify and execute an SQL statement from Show History by selecting the statement, which causes it to be displayed in the Execute Query text box. In Execute Query you can modify the SQL code and then click Execute. Making any change to an SQL statement retrieved from Show History causes it to be stored in Show History as a new statement; this include changes that do not affect execution, such as changing letter case, whitespace, or comments. Whitespace is not shown in Show History, but it is preserved when an SQL statement is retrieved from Show History.

You can execute (re-run) an unmodified SQL statement directly from the Show History list by clicking the Execute button found to the right of the SQL statement in the Show History listing.

You can also select buttons for Plan, Print, or Delete. The Print button displays the Print Query window, which gives you the options to either print or export to a file the query text and/or the query result set. The Delete button deletes the SQL statement from the history; the Delete All button at the end of the Show History list deletes all the SQL statements in the history.

Note that the Show History listing is not the same as the list of cached queries. Show History lists all invoked SQL statements from the current session, including those that failed during execution.

InterSystems IRIS supports numerous other ways to write and execute SQL code. These include:

• Embedded SQL: SQL code embedded within ObjectScript code.

• Dynamic SQL: using %SQL.StatementOpens in a new tab class methods (or other result set class methods) to execute SQL statements from within ObjectScript code.

• SQL Shell: executing Dynamic SQL from the Terminal using the SQL Shell interface.

The Browse tab provides a convenient way to quickly view all the schemas in a namespace, or a filtered subset of the schemas in the namespace. You can select Show All Schemas or Show Schemas with Filter, which applies the filter specified on the left side of the Management Portal SQL interface. By clicking on the Schema Name heading, you can list the schemas in ascending or descending alphabetical order.

Each listed schema provides links to lists of its associated Tables, Views, Procedures, and Queries (cached queries). If the schema has no items of that type, a hyphen (rather than a named link) is shown in that schema list column. This enables you to quickly get information about the contents of schemas.

Clicking a Tables, Views, Procedures, or Queries link displays a table of basic information about those items. By clicking on a table heading, you can sort the list by that column’s values in ascending or descending order. The Procedures table always includes Extent procedures, regardless of the Procedures setting on the left side of the Management Portal SQL interface.

You can get more information on individual Tables, Views, Procedures, and Cached Queries using the Catalog Details tab. Selecting a Table or View from the Browse tab does not activate the Open Table link for that table.

The following Catalog Details options are provided for each table:

• Table Info: Table Type: either TABLE, GLOBAL TEMPORARY, or SYSTEM TABLE (system tables are only displayed if the System check box is selected), Owner name, Last Compiled timestamp, External and Readonly boolean values, Class Name, Extent Size, the name of the Child Table(s) and/or the Parent Table (if relevant) and one or more References fields to other tables (if relevant), whether it uses the %Storage.Persistent default storage class, whether it Supports Bitmap Indexes, the RowID field name, a list of the fields that RowId is based on (if relevant), and whether the table is sharded. If there is an explicit shard key, it displays the shard key fields.

  Class Name is a link to the corresponding entry in the InterSystems Class Reference documentation. The Class Name is a unique package.class name derived from the table name by removing punctuation characters, as described in Identifiers and Class Entity Names.

  References only appears in Table Info if there is one or more references from a field in the current table to another table. These references to other tables are listed as links to the Table Info for the referenced table.

  Sharded: if the table is a shard-master table, the Table Info displays the name of the shard-local class and table with a link to the corresponding entry in the class documentation. If the table is a shard-local table, the Table Info displays the name of the shard-master class and table with a link to the corresponding entry in the class documentation. Shard-local tables are only displayed if the System check box is selected.

  This option also provides a modifiable value for the Number of rows to load when table is opened. This sets the maximum number of rows to display in Open Table. The available range is from 1 to 10,000; the default is 100. The Management Portal corrects a value outside the available range to a valid value: 0 corrects to 100; a fractional number rounds up to the next higher integer; a number greater than 10,000 corrects to 10,000.

• Fields: a list of the fields (columns) in the table showing: Field Name, Datatype, Column #, Required, Unique, Collation, Hidden, MaxLen, MaxVal, MinVal, Stream, Container, SQLType, Reference To, Version Column, Selectivity, Outlier Selectivity, and Average Field Size.

• Maps/Indices: a list of the indexes defined for the table showing: Index Name, SQL Map Name, Columns, Type, Block Count, Map Inherited, and Global.

  Index Name is the index property name and follows property naming conventions; when generated from an SQL index name, punctuation characters (such as underscores) in the SQL index name are stripped out. The SQL Map Name is the SQL name for the index. A generated SQL Map Name is the same as the Constraint Name, and follows the same naming conventions (described below). Columns specifies a field or a comma-separated list of fields specified for the index; it may specify the index collation type and full schema.table.field reference, as in the following example: $$SQLUPPER({Sample.People.Name}). Type can be one of the following: Bitmap Extent, Data/Master, Index (standard index), Bitmap, or Bitslice index, and the Unique constraint. The Block Count contains both the count and how that count was determined: set explicitly by the class author (Defined), computed by TuneTable (Measured), or estimated by the class compiler (Estimated). If Map Inherited? is Yes, this map was inherited from a superclass. Global is the name of the subscripted global containing the index data. The naming conventions for index globals are described in Index Global Names. You can supply this global name to ZWRITE to display the index data.

  This option also provides a link for each index to rebuild the index.

• Triggers: a list of the triggers defined for the table showing: Trigger Name, Time Event, Order, Code.

• Constraints: a list of the constraints for fields of the table showing: Constraint Name, Constraint Type, and Constraint Data (field name(s) listed in parentheses). Constraints include primary key, foreign key, and unique constraints. A primary key is, by definition, unique; it is only listed once. This option list constraints by constraint name; a constraint involving multiple fields is listed once with Constraint Data displaying a comma-separated list of the component fields. The Constraint Type can be UNIQUE, PRIMARY KEY, Implicit PRIMARY KEY, FOREIGN KEY, or Implicit FOREIGN KEY.

  You can also list constraints by invoking INFORMATION_SCHEMA.CONSTRAINT_COLUMN_USAGE. This list constraints by field name. The following example returns the name of the field and the name of the constraint for all UNIQUE, PRIMARY KEY, FOREIGN KEY and CHECK constraints:

  SELECT Column_Name,Constraint_Name FROM INFORMATION_SCHEMA.CONSTRAINT_COLUMN_USAGE WHERE TABLE_SCHEMA='Sample' AND TABLE_NAME='Person'

  

  If the table is defined with %PUBLICROWID and no explicit primary key is defined, the RowID field is listed with a Constraint Type of Implicit PRIMARY KEY with the Constraint Name RowIDField_As_PKey.

  For explicit constraints, the Constraint Name is generated as follows:

  • Constraint specified in the field definition: For example, FullName VARCHAR(48) UNIQUE or FullName VARCHAR(48) PRIMARY KEY. The Constraint Name value for the field is a generated value with the syntax TABLENAME_CTYPE#, where CTYPE is UNIQUE, PKEY, or FKEY, and # is a sequential integer assigned to unnamed constraints in the order specified in the table definition. For example, if FullName has the 2nd unnamed unique constraint (excluding the ID field) in the MyTest table, the generated Constraint Name for FullName would be MYTEST_UNIQUE2; if FullName is the primary key and the 3rd unnamed constraint (excluding the ID field) specified in the MyTest table, the generated Constraint Name for FullName would be MYTEST_PKEY3.

  • CONSTRAINT keyword named constraint clause: For example, CONSTRAINT UFullName UNIQUE(FirstName,LastName) or CONSTRAINT PKName PRIMARY KEY(FullName)), the Constraint Name is the specified unique constraint name. For example, FirstName and LastName in the MyTest table would each have the Constraint Name UFullName; FullName would have the Constraint Name PKName.

  • Unnamed constraint clause: For example, UNIQUE(FirstName,LastName) or PRIMARY KEY (FullName). The Constraint Name value is a generated value with the syntax TABLENAMECType#, where CType is Unique, PKey, or FKey, and # is a sequential integer assigned to unnamed constraints in the order specified in the table definition. For example, if FirstName and LastName have the 2nd unnamed unique constraint (excluding the ID field) in the MyTest table, the generated Constraint Name for FirstName and LastName would be MYTESTUnique2; if FullName is the primary key and the 3rd unnamed constraint (excluding the ID field) specified in the MyTest table, the generated Constraint Name for FullName would be MYTESTPKey3. (Note mixed uppercase/lowercase and absence of an underscore.)

  If a field is involved in more than one uniqueness constraint, it is listed separately for each Constraint Name.

• Cached Queries: a list of the cached queries for the table showing: Routine name, Query text, Creation Time, Source, Query Type.

• Table’s SQL Statements: a list of the SQL Statements generated for this table. Same information as namespace-wide SQL Statements display.

Management Portal SQL interface also provides Catalog Details for views, procedures, and cached queries:

The following Catalog Details options are provided for each view:

• View Info: Owner name, Last Compiled timestamp. This timestamp updates when you use the Edit View link and save changes.

  Defined as Read Only and View is Updateable booleans: if view definition included WITH READ ONLY, these are set to 1 and 0 respectively. Otherwise, if the view is defined from a single table they are set to 0 and 1; if the view is defined from joined tables they are set to 0 and 0. You can change this option using the Edit View link.

  Class Name is a unique package.class name derived from the view name by removing punctuation characters, as described in Identifiers and Class Entity Names.

  Check Option is only listed if the view definition included the WITH CHECK OPTION clause. It can be LOCAL or CASCADED. You can change this option using the Edit View link.

  Class Type is VIEW. It provides an Edit View link to edit the view definition.

  View Text is the SELECT statement used to define the view. You can change the view definition using the Edit View link.

  The list of fields includes the Field Name, Data Type, MAXLEN Parameter, MAXVAL Parameter, MINVAL Parameter, BLOB (%Stream.GlobalCharacter or %Stream.GlobalBinary field), Length, Precision, and Scale.

• View’s SQL Statements: a list of the SQL Statements generated for this view. Same information as namespace-wide SQL Statements display.

The following Catalog Details options are provided for each procedure:

• Stored Procedure Info:

  Class Name is a unique package.class name derived from the procedure name by pre-pending a type identifier ( ‘func’, ‘meth’, ‘proc’, or ‘query’) to the class name (for example, the SQL function MyProc becomes funcMyProc) and removing punctuation characters, as described in Identifiers and Class Entity Names. Class Document is a link to the corresponding entry in the class documentation. Procedure Type (for example, function). Method or Query Name the name of the generated class method or class query; this name is generated described in Identifiers and Class Entity Names. The Run Procedure link provides an option to interactively run the procedure.

• Stored Procedure’s SQL Statements: a list of the SQL Statements generated for this stored procedure. Same information as namespace-wide SQL Statements display.

Cached Query provides the full text of the query, an option to show the query execution plan, and an option to interactively execute the cached query.