Reader Directives

The directives that are processed by the Pervasive PSQL Reader are listed below. The suffixes shown are prefixed by the current <ReaderKeyword> in a mapping file. By default, the <ReaderKeyword> for the Pervasive PSQL Reader is PERVASIVE.

DATASET

Required/Optional: Required

This is the ODBC data source name.

Example:

PERVASIVE_DATASET citySource

Workbench Parameter: Source Pervasive PSQL DSN

DEF

Required/Optional: Required

The syntax of the definition is:

PERVASIVE_DEF <tableName> \
	[pervasive_where_clause	<whereClause>]	\
[<fieldName><fieldType>] +

OR

PERVASIVE_DEF <queryName> \
	[pervasive_sql_statement <sqlQuery>] \

The <tableName> must match the name of an existing Pervasive PSQL table in the database. This will be used as the feature type of all the features read from the table. The exception to this rule is when using the pervasive_sql_statement directive. In this case, the DEF name may be any valid alphabetic identifier; it does not have to be an existing table name – rather, it is an identifier for the custom SQL query. The feature type of all the features returned from the SQL query are given the query name.

The <fieldType> of each field must be given, but it is not verified against the database definition for the field. In effect, it is ignored.

The definition allows specification of separate search parameters for each table. If any of the per table configuration parameters are given, they will override, for that table, whatever global values have been specified by the reader directives such as the WHERE_CLAUSE. If any of these parameters is not specified, the global values will be used.

The following table summarizes the definition line configuration parameters:

Parameter

Contents

pervasive_where_clause

This specifies the SQL WHERE clause applied to the attributes of the layer’s features to limit the set of features returned. If this is not specified, then all the rows are returned. This directive will be ignored if the pervasive_sql_statement is present.

pervasive_sql_statement

This specifies an SQL SELECT query to be used as the source for the results. If this is specified, the Pervasive PSQL reader will execute the query, and use the resulting rows as the features instead of reading from the table <queryName>. All returned features will have a feature type of <queryName>, and attributes for all columns selected by the query. The pervasive_where_clause is ignored if pervasive_sql_statement is supplied. This form allows the results of complex joins to be returned to FME.

If no <whereClause> is specified, all rows in the table will be read and returned as individual features. If a <whereClause> is specified, only those rows that are selected by the clause will be read. Note that the <whereClause> does not include the word WHERE.

The Pervasive PSQL reader allows one to use the pervasive_sql_statement parameter to specify an arbitrary SQL SELECT query on the DEF line. If this is specified, FME will execute the query, and use each row of data returned from the query to define at least one feature. Each of these features will be given the feature type named in the DEF line, and will contain attributes for every column returned by the SELECT. In this case, all DEF line parameters regarding a WHERE clause or spatial querying are ignored, as it is possible to embed this information directly in the text of the <sqlQuery>.

In the following example, the all records whose ID is less than 5 will be read from the supplier table:

PERVASIVE_DEF supplier 	\
    pervasive_where_clause "id < 5" \
    ID  integer \
    NAME char(100) \
    CITY char(50) 

In this example, the results of joining the employee and city tables are returned. All attributes from the two tables will be present on each returned feature. The feature type will be set to complex.

PERVASIVE_DEF complex \
	pervasive_sql_statement	 \
		"SELECT * FROM EMPLOYEE, CITY WHERE EMPLOYEE.CITY = CITY.NAME"

WHERE_CLAUSE

Required/Optional

Optional

This optional specification is used to limit the rows read by the reader from each table. If a given table has no pervasive_where_clause or pervasive_sql_statement specified in its DEF line, the global <ReaderKeyword>_WHERE_CLAUSE value, if present, will be applied as the WHERE specifier of the query used to generate the results. If a table’s DEF line does contain its own pervasive_where_clause or pervasive_sql_statement, it will override the global WHERE clause.

The syntax for this clause is:

PERVASIVE_WHERE_CLAUSE <whereClause>

Note: The <whereClause> does not include the word “WHERE.”

The example below selects only the features whose lengths are more than 2000:

PERVASIVE_WHERE_CLAUSE LENGTH > 2000

Workbench Parameter

Where Clause

IDs

Required/Optional: Optional

This optional specification is used to limit the available and defined database tables that will be read. If no IDs are specified, then no tables are read. The syntax of the IDs directive is:

PERVASIVE_IDs <featureType1> \
<featureType2> … \
<featureTypeN> 

The feature types must match those used in DEF lines.

The example below selects only the HISTORY table for input during a translation:

PERVASIVE_IDs HISTORY

READ_CACHE_SIZE

Required/Optional: Optional

This directive controls how the reader retrieves rows from the database. This must be a numeric value which must be greater than 0.

The READ_CACHE_SIZE is used to determine the number of rows that are retrieved at one time into local memory from the data source. For example, if the READ_CACHE_SIZE is set to 10, after the reader is opened, the reader will read 10 rows into local memory. As features are processed by FME, the reader returns the data from the local memory buffer. As soon as you move past the last row available in local memory, the reader will retrieve the next 10 rows from the data source.

This directive affects the performance of the reader, and will result in significantly degraded performance if incorrectly set. The optimum value of this directive depends primarily on the characteristics of individual records and the transport between the database and the client machine. It is less affected by the quantity of rows that are to be retrieved.

By default, the READ_CACHE_SIZE is set to 10. This value has been determined to be the optimal value for average datasets.

Workbench Parameter: Number of Records to Fetch At A Time

RETRIEVE_ALL_SCHEMAS

This parameter is applicable only when generating a mapping file, generating a workspace, or when retrieving schemas in an FME Objects application.

When set to yes, schemas for all of the tables and feature classes in the database are returned.

When set to no (or missing), and if RETRIEVE_ALL_TABLE_NAMES is similarly set to no (or missing), only schemas requested by the IDs parameter are returned.

Required/Optional

Optional

Values

YES | NO (default)

NO: The reader will return the schemas for the feature types specified in the IDs. If no features are specified in IDs, then FME returns the schema features for all the tables.

If this value is not specified, then it is assumed to be No.

YES: Indicates to the reader to return all the schemas of the tables in the database.

Mapping File Syntax

Not applicable.

FME Objects applications would include RETRIEVE_ALL_SCHEMAS followed by “YES” in the parameters array passed to IFMEUniversalReader::open().

Workbench Parameter

Not applicable

RETRIEVE_ALL_TABLE_NAMES

This parameter is only applicable when generating a mapping file, generating a workspace or when retrieving schemas in an FME Objects application.

When set to yes, and if RETRIEVE_ALL_SCHEMAS is set to no (or missing), names for all of the tables and feature classes in the database are returned. When set to no (or missing), and if RETRIEVE_ALL_SCHEMAS is similarly set to no (or missing), the schemas requested by the IDs directive are returned.

Note: If RETRIEVE_ALL_SCHEMAS is also set to Yes, then RETRIEVE_ALL_SCHEMAS takes precedence.

Required/Optional

Optional

Values

YES | NO (default)

Mapping File Syntax

Not applicable.

FME Objects applications would include RETRIEVE_ALL_TABLE_NAMES followed by “YES” in the parameters array passed to IFMEUniversalReader::open().

Workbench Parameter

Not applicable (used when you browse a Table List)

EXPOSED_ATTRS

This directive allows the selection of format attributes to be explicitly added to the reader feature type.

This is similar to exposing format attributes on a reader feature type once it has been generated; however, it is even more powerful because it enables schema-driven applications other than Workbench to access and leverage these attributes as if they were explicitly on the schema as user attributes.

The result of picking a list of attributes is a comma-separated list of attribute names and types that will be added to the schema features. Currently all reader feature types will receive the same set of additional schema attributes for a given instance of the reader.

Required/Optional

Optional

Mapping File Syntax

Not applicable.

While it is possible for FME Objects applications to invoke this directive, the required format is not documented.

This directive is intended for use in our GUI applications (for example, Workbench) only.

Workbench Parameter

Additional Attributes to Expose