DBExecute

Short Description

DBExecute executes specified SQL/DML/DDL statements against a database connected using the JDBC driver. It can execute queries, transactions, call stored procedures or functions.

Input parameters can be received through the single input port and output parameters or result set are sent to the first output port. Error information can be sent to the second output port.

Hint!

For executing INSERT/DELETE/UPDATE SQL queries operating on single records read from input port of the component, consider using DatabaseWriter instead, as it is optimized for these types of queries.

Ports

[1] Input port must be connected if the Query input parameters attribute is specified or if the whole SQL query is received through the input port.

[2] The output port must be connected if the Query output parameters or the Return set output fields attribute is specified.

Metadata

DBExecute propagates input metadata to the first output port if not calling a stored procedure. It adds two output fields: updateCount and statement.

DBExecute has no metadata template.

Metadata on both output ports may contain any number of fields from input (same names and types). Input metadata are mapped automatically according to their name(s) and type(s).

Standard output metadata may contain two additional fields: updateCount and statement (case insensitive, so UpdateCount or UPDATE_COUNT also works). If present, these fields are populated with the number of updated rows and the executed statement, respectively. This feature only works if the Call as stored procedure attribute is set to false.

Error output metadata may contain additional fields for error information. The two error fields may have any names and must be set to the following Autofilling Functions: ErrCode and ErrText.

DBExecute Attributes

[1] One of these must be set. If both are specified, Query URL has higher priority.

Details

SQL Query Received from Input Port

SQL query can also be received from the input port.

In this case, two values of the Query URL attribute are allowed:

• SQL command is sent through the input edge.
  The attribute value is: port:$0.fieldName:discrete.
  Metadata of this edge has neither default delimiter, nor record delimiter, but EOF as delimiter must be set to true.

• The name of the file containing the SQL command, including the path, is sent through the input edge.
  The attribute value is: port:$0.fieldName:source.

For more details about reading data from input port, see Input Port Reading.

Transaction Set

Options are the following:

• One statement
  The commit is performed after each query execution.

• One set of statements
  All statements are executed for each input record. The commit is performed after a set of statements.
  For this reason, if an error occurs during the execution of any statement for any of the records, all statements are rolled back for such a record.

• All statements
  The commit is performed after all statements, only.
  For this reason, if an error occurs, all operations are rolled back.

• Never commit
  The commit or rollback may be called from other component in a different phase. There is no automatic rollback.

Warning!

If no error occurs, the connection closure results in autocommit even if Never commit is selected. If you need to rollback, the rollback must be called before autocommit on session’s termination.

If you want to use the Never commit option and perform commit or rollback from another component in another phase, set Thread-safe connection in advanced Connection settings to false. Otherwise, each component will have different connection and autocommit will be performed at the end of processing of particular components.

Calling Stored Procedures and Functions

The following table summarizes basic configuration examples to call stored procedures, setting input parameters and obtaining output values in the form of output parameters, return values or cursors.

Examples

Creating a database table with DBExecute

This example shows a basic use case of this component.

Create a database table rivers with two columns: name and length.

Solution

Create a database connection RiversConnection. See Creating Internal Database Connections.

Set the DB Connection to RiversConnection connection.

Enter the CREATE statement into the SQL Query attribute.

1. ATTRIBUTE
   DB Connection
   VALUE
   RiversConnection

2. ATTRIBUTE
   Query URL
   VALUE

CREATE TABLE rivers (
    name VARCHAR(50) NOT NULL,
    length INTEGER
);

If you use one graph to create a database table and insert data to it, do no forget to put DBExecute and the component inserting the data into different phases. DBExecute should be in a lower phase and DatabaseWriter (or other component writing to the database) in a higher phase.

Executing multiple queries

DBExecute is not limited to execution of one SQL query.

Create table rivers as in the previous example. If the table exists, drop it first.

Solution

1. ATTRIBUTE
   DB Connection
   VALUE
   RiversConnection

2. ATTRIBUTE
   Query URL
   VALUE

DROP TABLE IF EXISTS rivers;
CREATE TABLE rivers (
    name VARCHAR(50) NOT NULL,
    length INTEGER
);

Do not forget to end statements with semicolons.

Note: some databases (e.g. Oracle) do not understand the IF EXISTS part of the query. If you use a database that does not support IF EXISTS, change the first line of the query to DROP TABLE rivers;.

Creating a stored procedure

Create a stored procedure that inserts data into the products table. The table has been created with the following query:

CREATE TABLE products (
    product_name VARCHAR(40) NOT NULL
);

Solution

Create a database connection ProductConnection.

1. ATTRIBUTE
   DB Connection
   VALUE
   ProductConnection

2. ATTRIBUTE
   Query URL
   VALUE

CREATE OR REPLACE FUNCTION add_product(param VARCHAR(40))
RETURNS VOID AS '
BEGIN
    INSERT INTO products (product_name) VALUES(param);
END;
' LANGUAGE plpgsql;

The example has been tested with PostgreSQL 9.2 and 14.4.

Getting errors

This example shows a way to get errors related to an SQL query from the DBExecute component to send it to the next component for further processing.

The graph uses DBExecute to create a database table bricks. The bricks table contains details on products from our supplier. The table might exist or we could not have permission to create a table and we would like to receive the errors from the database.

Solution

Create a database connection ProductConnection.

1. ATTRIBUTE
   DB Connection
   VALUE
   ProductConnection

2. ATTRIBUTE
   Query URL
   VALUE

CREATE TABLE bricks (
    id INTEGER PRIMARY KEY NOT NULL,
    length integer,
    width integer,
    height integer,
    weight integer
);

Connect an edge to the second output port of DBExecute.

Assign metadata to the edge. The metadata should contain a String field that has Autofilling set to ErrCode.

Best Practices

Upload and Download of Data

Do not use the DBExecute component for INSERT and SELECT statements. For uploading data to a database, please use the DatabaseWriter component. Similarly, for downloading use the DatabaseReader component.

Transferring data within a database

The best practice to load data from one table to another in the same database is to do it inside the database. You can use the DBExecute component with a query like this:

insert into my_table select * from another_table

because pulling data out from the database and putting them in is slower as the data has to be parsed during the reading and formatted when writing.

Charset

If the query is specified in an external file (with Query URL), we recommend users to explicitly specify Query source charset.

Troubleshooting

Boolean and Oracle JDBC

Calling arguments or return values of the PL/SQL RECORD, BOOLEAN, or table with non-scalar elements are not supported by Oracle JDBC Drivers. See Oracle JDBC Reference

As a workaround, you can create a wrapper procedure. See Wrapper procedures

See also
DatabaseReader
DatabaseWriter
Common properties of components
Specific attribute types