Data Types (SQL)

Literal Parameters

Literal parameters are identified in the DDL data type and the InterSystems IRIS data type in the format %n, where n is the number of the data type argument. For example, VARCHAR(%1) maps to %String(MAXLEN=%1)

Common literal parameters include maximum string length, minimum and maximum values, and precision and scale values.

Maximum Length

In data type classes, the MAXLEN parameter specifies the maximum length of string data types. DDL types often define these values in a corresponding unnamed parameter.

In this field definition, the data type is a string with a maximum length of 64 characters.

DDL Using SQL Data Type DDL Using InterSystems IRIS Data Type

ProductName VARCHAR(64)

ProductName %String(MAXLEN=64)

When specifying this parameter, keep these points in mind:

• A field with no MAXLEN value can take a value of any length, up to the maximum string length. To define a string field of maximum length, specify VARCHAR(''), which create a property with data type %Library.String(MAXLEN=""). VARCHAR() creates a property with data type %Library.String(MAXLEN=1). To define a binary field with no MAXLEN value, specify VARBINARY(''), which create a property with data type %Library.Binary(MAXLEN=""). VARBINARY() creates a property with data type %Library.Binary(MAXLEN=1).

• Large MAXLEN: ODBC applications may be affected by an overly large MAXLEN value. ODBC applications try to make decisions about the size of a field needed based on metadata from the server, so the application may allocate more buffer space than is actually needed. For this reason, InterSystems IRIS supplies a system-wide default ODBC VARCHAR maximum length of 4096; this system-wide default is configurable using the Management Portal: from System Administration, select Configuration, then SQL and Object Settings, then SQL. View or set the Default length for VARCHAR option. To determine the current setting, call $SYSTEM.SQL.CurrentSettings()Opens in a new tab. The InterSystems ODBC driver takes the data from the TCP buffer and converts it into the applications buffer, so MAXLEN size does not affect our ODBC client.

Maximum and Minimum Values

In data type classes, the MINVAL and MAXVAL parameters specify the minimum and maximum values of numeric data types. Data type classes often define these values using function parameters instead of literal parameters. DDL types do not have equivalent parameters.

In this field definition, the data type is an integer from 0 to 100.

Capacity %Integer(MINVAL=0,MAXVAL=100)

Precision and Scale

In data type classes, the PRECISION and SCALE parameters are integer values specifying the precision (maximum number of digits) and scale (maximum number of decimal digits) of numeric data types. Data type classes often define these values as function parameters instead of literal parameters. DDL data types such as NUMERIC often specify precision and scale together as unnamed parameters.

This field definition defines a number that has a precision of 6 and a scale of 2.

UnitPrice NUMERIC(6,2) // Range: -9999.99 to 9999.99

The precision and scale parameters define numeric data types as follows:

• Precision — The maximum and minimum permitted value, specified as an integer from 0 to 19 + s, where s is the scale. Precision is commonly the total number of digits in the number, but its exact value is determined by the %Library class data type mapping. The maximum integer value is 9223372036854775807. A precision larger than 19 + s defaults to 19 + s.

• Scale — The maximum number of decimal (fractional) digits permitted, specified as an integer. If s is larger than or equal to the precision, p, only a fractional value is permitted, and the actual p value is ignored. The largest permitted scale is 18, which corresponds to .999999999999999999. A scale larger than 18 defaults to 18.

For information about numeric formatting, refer to the $FNUMBER function.

Function Parameters

Function parameters are used when a parameter in the DDL data type parameter must be transformed before it can be put into the InterSystems IRIS data type. An example of this is the transformation of a DDL data type’s numeric precision and scale parameters into an InterSystems IRIS data type’s MAXVAL, MINVAL parameters.

For example, consider the mapping between the DECIMAL DDL data type and the %NumericOpens in a new tab class as it appears in the System-defined DDL Mappings table in the Management Portal.

DDL Using SQL Data Type DDL Using InterSystems IRIS Data Type

DECIMAL(%1,%2)

%Numeric(MAXVAL=<|'$$maxval^%apiSQL(%1,%2)'|>,MINVAL=<|'$$minval^%apiSQL(%1,%2)'|>,SCALE=%2)

The %1 and %2 parameters specify the precision and scale of numbers in that data type, respectively. For example, a field of type DECIMAL(4,2) stores numbers using a precision of 4 and a scale of 2. InterSystems SQL uses these parameters to derive the minimum value (–99.99) and maximum value (99.99) accepted by the field.

For the %NumericOpens in a new tab class, InterSystems IRIS set the SCALE parameter (SCALE=%2) but it does not have a PRECISION parameter to set. Instead, InterSystems IRIS sets the MAXVAL and MINVAL parameters using these transformation functions:

• maxval^%apiSQL(precision,scale) returns the maximum valid numeric value, MAXVAL, given the precision and scale.

• minval^%apiSQL(precision,scale) returns the minimum valid numeric value, MINVAL, given the precision and scale.

The syntax for these transformation functions is as follows:

dataTypeClass(param=<|'func'|>, param2=<|func2|>, ...)

Additional Parameters

A data type class may define additional data definition parameters that cannot be defined using a DDL data type. These include data validation operations such as an enumerated list of permitted data values, pattern matching of permitted data values, and automatic truncation of data values that exceed the MAXLEN maximum length.

Date, Time, and Timestamp Data Types

Using standard InterSystems SQL date, time functions, you can define date, time, and timestamp data types. You can also convert between dates and timestamps. For example, you can use CURRENT_DATE or CURRENT_TIMESTAMP as input to a field defined with that data type, or use DATEADD, DATEDIFF, DATENAME, or DATEPART to manipulate date values stored with this data type.

This table shows how date, time, and timestamp data type classes map to SQL types. The data type classes use this type when performing calculations in SQL. When creating a custom data type class, you can use these mappings to determine which SQL type to specify in the SqlCategory keyword of your class definition. For example:

Class MyApp.MyDateDT [ ClassType = DataType, SQLCategory = DATE ]
{
    // class members
}

You can compare POSIXTIME to DATE or TIMESTAMP values using =, <>, >, or < operators. Refer to Overview of Predicates for further details.

When comparing FMTIMESTAMP category values with DATE category values, InterSystems IRIS does not strip the time from the FMTIMESTAMP value before comparing it to the DATE. This is identical behavior to comparing TIMESTAMP with DATE values, and comparing TIMESTAMP with MVDATE values. It is also compatible with how other SQL vendors compare timestamps and dates. This means a comparison of a FMTIMESTAMP 320110202.12 and DATE 62124 are equal when compared using the SQL equality (=) operator. Applications must convert the FMTIMESTAMP value to a DATE or FMDATE value to compare only the date portions of the values.

Dates Prior to December 31, 1840

A date is commonly represented by the DATE data type or the TIMESTAMP data type.

The DATE data type stores a date in $HOROLOG format, as a positive integer count of days from the arbitrary starting date of December 31, 1840. By default, dates can only be represented by a positive integer (MINVAL=0), which corresponds to the date December 31, 1840. However, you can change the %Library.DateOpens in a new tab MINVAL type parameter to enable storage of dates prior to December 31, 1840. By setting MINVAL to a negative number, you can store dates prior to December 31, 1840 as negative integers. The earliest allowed MINVAL value is -672045. This corresponds to January 1 of Year 1 (CE). DATE data type cannot represent BCE (also known as BC) dates.

The TIMESTAMP data type defaults to 1840–12–31 00:00:00 as the earliest allowed timestamp. However, you can change the MINVAL parameter to define a field or property that can store dates prior to December 31, 1840. For example, MyTS %Library.TimeStamp(MINVAL='1492-01-01 00:00:00'). The earliest allowed MINVAL value is 0001–01–01 00:00:00. This corresponds to January 1 of Year 1 (CE). The %TimeStamp data type cannot represent BCE (also known as BC) dates.

Note:

Be aware that these date counts do not take into account changes in date caused by the Gregorian calendar reform (enacted October 15, 1582, but not adopted in Britain and its colonies until 1752).

You can redefine the minimum date for your locale as follows:

  SET oldMinDate = ##class(%SYS.NLS.Format).GetFormatItem("DATEMINIMUM")
  IF oldMinDate=0 {
    DO ##class(%SYS.NLS.Format).SetFormatItem("DATEMINIMUM",-672045)
    SET newMinDate = ##class(%SYS.NLS.Format).GetFormatItem("DATEMINIMUM")
    WRITE "Changed earliest date to ",newMinDate
    }
  ELSE { WRITE "Earliest date was already reset to ",oldMinDate}

The above example sets the MINVAL for your locale to the earliest permitted date (1/1/01). For more details on configuring dates based on your locale, see Configuring National Language Support (NLS).

Note:

InterSystems IRIS does not support using Julian dates with negative logical DATE values (%Library.Date values with MINVAL<0). Thus, these MINVAL<0 values are not compatible with the Julian date format returned by the TO_CHAR function.

Strings

InterSystems IRIS permits a fixed amount of memory to handle strings, so that there is a string length limit. Commonly, extremely long strings should be assigned one of the %Stream.GlobalCharacterOpens in a new tab data types.

No string length limit is enforced over a database driver connection. If the InterSystems IRIS instance and the ODBC driver facilities support different protocols, the lower of the two protocols is used. The protocol that was actually used is recorded in the InterSystems ODBC log.

Note that, by default, InterSystems IRIS establishes a system-wide ODBC VARCHAR maximum length of 4096; this ODBC maximum length is configurable.

List Structures

InterSystems IRIS supports the list structure data type %List (data type class %Library.List). This is a compressed binary format, which does not map to a corresponding native data type for InterSystems SQL. In its internal representation it corresponds to data type VARBINARY with a default MAXLEN of 32749. InterSystems IRIS supports the list structure data type %ListOfBinary (data type class %Library.ListOfBinary) corresponds to data type VARBINARY with a default MAXLEN of 4096.

For this reason, Dynamic SQL cannot use %List data in a WHERE clause comparison. You also cannot use INSERT or UPDATE to set a property value of type %List.

Dynamic SQL returns the data type of list structured data as VARCHAR. To determine if a field in a query is of data type %List or %ListOfBinary you can use the select-item columns metadata isList boolean flag. The CType (client data type) integer code for these data types is 6.

If you use an ODBC or JDBC client, %List data is projected to VARCHAR string data, using LogicalToOdbc conversion. A list is projected as a string with its elements delimited by commas. Data of this type can be used in a WHERE clause, and in INSERT and UPDATE statements. Note that, by default, InterSystems IRIS establishes a system-wide ODBC VARCHAR maximum length of 4096; this ODBC maximum length is configurable.

For further details on data type class %Library.ListOpens in a new tab, refer to the InterSystems Class Reference. For further details on using lists in a WHERE clause, see the %INLIST predicate and the FOR SOME %ELEMENT predicate. For further details on handling list data as a string, see the %EXTERNAL function.

InterSystems SQL supports eight list functions: $LIST, $LISTBUILD, $LISTDATA, $LISTFIND, $LISTFROMSTRING, $LISTGET, $LISTLENGTH, and $LISTTOSTRING. ObjectScript supports three additional list functions: $LISTVALID to determine if an expression is a list, $LISTSAME to compare two lists, and $LISTNEXT to sequentially retrieve elements from a list.

BIT Data Type

The BIT (%Library.BooleanOpens in a new tab) data type accepts 0, 1, and NULL as valid values.

• In Logical and ODBC modes the only accepted values are 0, 1, and NULL.

• In Display mode the DisplayToLogical method first translates a non-null input value to 0 or 1, as follows:

  • Non-zero numbers or numeric strings = 1. For example, 3, '0.1', '-1', '7dwarves'.

  • Non-numeric strings = 0. For example, 'true' or 'false'.

  • Empty string = 0. For example, ''.

Stream Data Types

The Stream data types correspond to the InterSystems IRIS class property data types %Stream.GlobalCharacterOpens in a new tab (for CLOBs) and %Stream.GlobalBinaryOpens in a new tab (for BLOBs). These data type classes can define a stream field with a specified LOCATION parameter, or omit this parameter and default to a system-defined storage location.

A field with a Stream data type cannot be used as an argument to most SQL scalar, aggregate, or unary functions. Attempting to do so generates an SQLCODE -37 error code. The few functions that are exceptions are listed in the Storing and Using Stream Data (BLOBs and CLOBs) chapter of Using InterSystems SQL.

A field with a Stream data type cannot be used as an argument to most SQL predicate conditions. Attempting to do so generates an SQLCODE -313 error code. The predicates that accept a stream field are listed in the Storing and Using Stream Data (BLOBs and CLOBs) chapter of Using InterSystems SQL.

A sharded table cannot contain stream data type fields.

The use of Stream data types in indices, and when performing inserts and updates are also restricted. For further details on Stream restrictions, refer to the Storing and Using Stream Data (BLOBs and CLOBs) chapter of Using InterSystems SQL.

SERIAL Data Type

A field with a SERIAL (%Library.CounterOpens in a new tab) data type can take a user-specified positive integer value, or InterSystems IRIS can assign it a sequential positive integer value. %Library.CounterOpens in a new tab extends %Library.BigIntOpens in a new tab.

An INSERT operation specifies one of the following values for a SERIAL field:

• No value, 0 (zero), or a nonnumeric value: InterSystems IRIS ignores the specified value, and instead increments this field's current serial counter value by 1, and inserts the resulting integer into the field.

• A positive integer value: InterSystems IRIS inserts the user-specified value into the field, and changes the serial counter value for this field to this integer value.

Thus a SERIAL field contains a series incremental integer values. These values are not necessarily continuous or unique. For example, the following is a valid series of values for a SERIAL field: 1, 2, 3, 17, 18, 25, 25, 26, 27. Sequential integers are either InterSystems IRIS-generated or user-supplied; nonsequential integers are user-supplied. If you wish SERIAL field values to be unique, you must apply a UNIQUE constraint on the field.

An UPDATE operation has no effect on automatically-assigned SERIAL counter field values. However, an update performed using INSERT OR UPDATE causes a skip in integer sequence for subsequent insert operations for a SERIAL field.

An UPDATE operation can only change a serial field value if the field currently has no value (NULL), or its value is 0. Otherwise, an SQLCODE -105 error is generated.

InterSystems IRIS imposes no restriction on the number of SERIAL fields in a table.

ROWVERSION Data Type

The ROWVERSION data type defines a read-only field that contains a unique system-assigned positive integer, beginning with 1. InterSystems IRIS assigns sequential integers as part of each insert, update, or %Save operation. These values are not user-modifiable.

InterSystems IRIS maintains a single row version counter namespace-wide. All tables in a namespace that contain a ROWVERSION field share the same row version counter. Thus, the ROWVERSION field provides row-level version control, allowing you to determine the order in which changes were made to rows in one or more tables in a namespace.

You can only specify one field of ROWVERSION data type per table.

The ROWVERSION field should not be included in a unique key or primary key. The ROWVERSION field cannot be part of an IDKey index.

For details on using ROWVERSION, refer to RowVersion Field section of the “Defining Tables” chapter of Using InterSystems SQL.

ROWVERSION and SERIAL Counters

Both ROWVERSION and SERIAL (%Library.CounterOpens in a new tab) data type fields receive a sequential integer from an internal counter as part of an INSERT operation. But these two counters are significantly different and are used for different purposes:

• The ROWVERSION counter is at the namespace level. The SERIAL counter is at the table level. These two counters are completely independent of each other and independent of the RowID counter.

• The ROWVERSION counter is incremented by insert, update, or %Save operations. The SERIAL counter is only incremented by insert operations. An update performed using INSERT OR UPDATE can cause a gap in the SERIAL counter sequence.

• A ROWVERSION field value cannot be user-specified; the value is always supplied from the ROWVERSION counter. A SERIAL field value is supplied from the table’s internal counter during an insert if you do not specify a value for this field. If an insert supplies a SERIAL integer value, that value is inserted rather than the current counter value:

  • If an insert supplies a SERIAL field value greater than the current internal counter value, InterSystems IRIS inserts that value into the field and resets the internal counter to that value.

  • If an insert supplies a SERIAL field value less than the current counter value, InterSystems IRIS does not reset the internal counter.

  • An insert can supply a SERIAL field value as a negative integer or a fractional number. InterSystems IRIS truncates a fractional number to its integer component. If the supplied SERIAL field value is 0 or NULL, InterSystems IRIS ignores the user-supplied value and inserts the current internal counter value.

  You cannot update an existing SERIAL field value.

• A ROWVERSION field value is always unique. Because you can insert a user-specified SERIAL field value, you must specify a UNIQUE field constraint to guarantee unique SERIAL field values.

• The ROWVERSION counter cannot be reset. A TRUNCATE TABLE resets the SERIAL counter; performing a DELETE on all rows does not reset the SERIAL counter.

• Only one ROWVERSION field is allowed per table. You can specify multiple SERIAL fields in a table.

DDL Data Types Exposed by InterSystems ODBC / JDBC

InterSystems ODBC exposes a subset of the DDL data types, and maps other data types to this subset of data types. These mappings are not reversible. For example, the statement CREATE TABLE mytable (f1 BINARY) creates an InterSystems IRIS class that is projected to ODBC as mytable (f1 VARBINARY). An InterSystems IRIS list data type is projected to ODBC as a VARCHAR string.

ODBC exposes the following data types: BIGINT, BIT, DATE, DOUBLE, GUID, INTEGER, LONGVARBINARY, LONGVARCHAR, NUMERIC, OREF, POSIXTIME, SMALLINT, TIME, TIMESTAMP, TINYINT, VARBINARY, VARCHAR. Note that, by default, InterSystems IRIS establishes a system-wide ODBC VARCHAR maximum length of 4096; this ODBC maximum length is configurable.

When one of these ODBC/JDBC data type values is mapped to InterSystems SQL, the following operations occur: DOUBLE data is cast using $DOUBLE. NUMERIC data is cast using $DECIMAL.

The GUID data type corresponds to InterSystems SQL UNIQUEIDENTIFIER data type. Failing to specify a valid value to a GUID / UNIQUEIDENTIFIER field generates a #7212 General Error. To generate a GUID value, use the %SYSTEM.Util.CreateGUID()Opens in a new tab method.