Home

	Expand Search

Open Client Client-Library/C Reference Manual

Chapter 2 Topics

Commands

Sending commands

Sending commands

All commands are defined and sent in three steps:

1. Initiate the command. This identifies the command type and what it executes.

2. Define parameter values, if necessary.

3. Send the command. ct_send writes the command symbols and data to the network. The server then reads the command, interprets it, and executes it.

An application sends several types of commands to a server:

• An application calls ct_command to initiate a language, message, package, remote procedure call (RPC), or send-data command.

• An application calls ct_cursor to initiate a cursor command.

• An application calls ct_dynamic to initiate a dynamic SQL command.

The following types of commands take parameters:

• A language command, when the command text contains variables

• An RPC command, when the stored procedure takes parameters

• A cursor declare command, when the body of the cursor contains host variables or when some of the cursor's columns are for update

• A cursor open command, when the body of the cursor contains host language parameters

• A cursor update command if the text of the update statement contains variables

• A message command

• A dynamic SQL execute command

An application calls ct_param or ct_setparam once for each parameter a command requires. These routines perform the same function, except that ct_param copies a parameter value, while ct_setparam copies the address of a variable that contains the value. If ct_setparam is used, Client-Library reads the parameter value when the command is sent. This allows the application to change the parameter values that were specified with ct_setparam before resending the command.

After a command has been initiated and its parameters have been defined, an application calls ct_send to send the command to the server. The server then interprets the command, executes it, and returns the results to the client application.

For most command types, Client-Library allows an application to resend the command after the results of previous execution have been processed. Enhancements to ct_send, ct_cursor, and ct_bind, and the addition of ct_setparam routine in version 11.1 allow batch-processing applications to resend commands and reuse binds when repeatedly executing the same server command. This feature can eliminate redundant calls to ct_bind, ct_command, ct_cursor, and ct_param.

The application resends commands as follows:

• If necessary, the application changes values in parameter source variables.

  If the command requires parameters, the application should define parameter source variables with ct_setparam instead of passing values with ct_param. Input parameter values passed with ct_param can not be changed when a command is resent.

• The application calls ct_send to resend the command after the results of the previous command execution have been processed and before a new command is initiated on the command structure.

An application can resend all types of commands except:

• Send-data commands initiated by ct_command(CS_SEND_DATA_CMD)

• Send-bulk commands initiated by ct_command(CS_SEND_BULK_CMD)

• ct_cursor cursor commands other than cursor-update or cursor-delete

• ct_dynamic commands other than execute-immediate commands or a command to execute a prepared statement