{{ extract authorization-identifier.table-identifier({column-name[{{,column-
name }} ...]|* })
   to update-information-name
    key|ukey(column-name[{{,column-name }} ...])[check {not_null_unique|unique|none}] }} ...

• extract authorization-identifier.table-identifier({column-name[{{,column-name }} ...]|* })
  Specify the table that is to be subject to extraction and the columns to be extracted from that table. You must specify one extract statement for each extraction condition. You can specify only one authorization-identifier.table-identifier per extract statement.
  To use multiple extraction conditions to extract a single table, specify a separate extract statement for each extraction condition. You can specify a maximum of 4000 column names per extract statement.
  

  authorization-identifier.table-identifier

  Specify the authorization identifier and table identifier of the table that contains the columns to be extracted. This must be a base table.

  authorization-identifier

  Specify the authorization identifier for the table that is to be subject to extraction of update information.

  table-identifier

  Specify the table identifier of the table that is to be subject to extraction of update information.

  column-name

  Specify the name of a column to be extracted. You can specify a maximum of 4000 columns. You can specify a repetition column only if no elements are specified, but array columns are not supported. You can specify the same column name more than once; however, if the column names are also specified in the key clause, they must be unique. The update information fields will reflect the order in which the columns were specified in the extract statement. An update information field is an area in the update information that is used to store an extracted HiRDB field.

  #

  Specify the asterisk (*) to have all columns in the specified table extracted as is without changing their order.

• to update-information-name
  ~ <symbolic name of 1-8 characters>
  Specify a name for the extracted update information. Each update information name must be unique among all the extraction statements in the extraction definition.
  
• {key|ukey}(column-name[{{,column-name }} ...])
  Specify the names of the columns to be used as the mapping key. Whether you use the key or ukey clause depends on whether the specified columns can be updated.
  For details about the data types of columns that can be used as mapping keys, see 4.3.3(1) Mapping key when the source database is HiRDB.
  

  key column-name

  Use the specified column as a mapping key column; this column cannot be updated.

  ukey column-name

  Use the specified column as a mapping key column; this column can be updated.

  Note that if a mapping key is updated, the data linkage recovery facility cannot be used.

  You can specify a maximum of 16 column names, but the same column name cannot be specified more than once. The specified column names must also be specified in the extract statement. If you specified * in the extract statement, you must specify one of the columns of the table that is to be subject to extraction processing.
  
• check {not_null_unique|unique|none}
  Specify the condition for performing a unique check on the mapping key column. The table below provides the details of unique checking. If the condition is not satisfied, extraction definition preprocessing results in a definition error. If this clause is omitted, Datareplicator performs checking based on the -k option specification in the hdeprep command.
  

  Table 5-7 Details of unique checking

  Check item	Description
  Index type	The index type must be one of the following: Unique index Unique cluster index Primary index Primary cluster index
  Index component column	There must be only the mapping key component columns.

  not_null_unique

  Check the source table to confirm that the index satisfying the conditions described in Table 5-7 Details of unique checking has been defined and its index component columns have the NOT NULL attribute.

  unique

  Check the source table to confirm that the index satisfying the conditions described in Table 5-7 Details of unique checking has been defined. Because this option does not check the NULL value, you need to use a tool such as a user program that updates the source database to ensure the uniqueness of the data.

  none

  Do not perform a unique check. You need to use a tool such as a user program that updates the source database to ensure the uniqueness of the data.

• If you specify the key clause, also specify columns that are not a mapping key in the extract statement. Because mapping keys with the key clause specified are not updated, data cannot be extracted if only mapping-key columns are specified. If you specify the ukey clause, you can specify mapping-key columns only in the extract statement.
• You can extract null-value data.
• The authorization identifier, table identifier, column name, update information name, and target identifier specifications are case-sensitive.
• If you specify an authorization identifier, table identifier, column name, update information name, or target identifier that is the same as a source Datareplicator reserved word, you must enclose it in double quotation marks ("). For details about the source Datareplicator reserved words, see Appendix B. Datareplicator Reserved Words.
• A column name specified as part of the mapping key cannot be specified more than once in the extract statement.
• If you update the mapping key when the key clause is specified, the target system uses the updated data as the key. For this reason, an error might occur during import processing (no data with matching key) or data might be imported into an unexpected row. It is not advisable to update the mapping key.