6.3.2 Details of API trace information

API trace information is provided as a trace file to which query entry and exit information is output. You can use this information to determine the startup and termination of adaptors and queries.

You can use the sdptrced command to edit and output the API trace information. The figure below shows an example of trace information that has been output as a result of executing the sdptrced command. This example outputs trace information on queries Q1 through Q3 that are in a mutually dependent relationship (Q1's output tuples are input to Q2, and Q2's output tuples are input to Q3).

Figure 6-1 Example output of API trace information

[Figure]

Notes:
In this figure, the version information for the Stream Data Platform - AF that output the trace file, the trace file type (API), and the path (absolute path) of the API trace file are output at the top.
A 30-digit value is set for each of trace-serial-number, reserved-area, and root-trace-serial-number.

From this example, you can obtain the following information (the numbers below correspond to the numbers on the right side of the figure):

  1. The value of Data is Q1 and the value of EventID is 0x00020000 (input of a tuple to the query).
    This trace information means that a tuple was input to query Q1 at the indicated time.
  2. The value of Data is Q1 and the value of EventID is 0x00020001 (output of a tuple from the query).
    This trace information means that a tuple was output from query Q1 at the indicated time.

The information in EventID and Data on the subsequent lines shows that input to and output from query Q2 and input to and output from query Q3 were processed in this order.

You can obtain the time of an input or output operation in nanoseconds from the value of nSec.

For details about the sdptrced command, see sdptrced (edits trace information) in 7. Commands.