5.1.12 Data Registration

Overview

This section describes the data registration for TimeSeries containers

Sample Data

Similar to collection sample, data to be registered is initially stored in a CSV file. Data must be registered in order to later read. The contents of the CSV file (Instrument_log.csv) is as follows:

• Data held
• Measuring instrument ID
• Timestamp
• Temperature
• Live image file path

List.1 Measurement Log Contents (instrument_log.csv)

weather_station_1,2016/7/1 0: 00,50, liveimage1.jpg
weather_station_2,2016/7/1 0: 00,50,
weather_station_3,2016/7/1 0: 00,50, liveimage2.jpg
weather_station_4,2016/7/1 0: 00,50,
weather_station_5,2016/7/1 0: 00,50, liveimage1.jpg
(Snip)

Data Registration Process

The following snippet demonstrates data registration.

List.2 Register TimeSeries Container (TimeSeriesRegister.java)

// Read InstrumentLog data from csv
List<InstrumentLog> logList = logLogic.readCsv();
// Register TimeSeries Container
logLogic.registerTimeSeriesContainer(store, logList);

• L.32: Process the CSV into a list of Instrument Logs.
• L.34: Register the parsed data in GridDB.

List.3 Read Measurement Log (InstrumentLogLogic.java)

public List<InstrumentLog> readCsv()
    throws IOException, ParseException, SerialException, SQLException {
    // Read CSV file
    List<InstrumentLog> logList = new ArrayList<InstrumentLog>();
    SimpleDateFormat format = new SimpleDateFormat("yyyy/MM/dd HH: mm", Locale.US);
    String[] line = null;
    CSVReader reader = new CSVReader(new FileReader("data/instrument_log.csv"));

    try {
        while ((line = reader.readNext()) != null) {
            InstrumentLog log = new InstrumentLog();
            log.weatherStationId = line[0];
            log.timestamp = format.parse(line[1]);
            log.temperture = Float.valueOf(line[2]);
    
            // Write Blob data
            ByteArrayOutputStream outputStream = new ByteArrayOutputStream();
            String filePath = line[3];
            if ((filePath != null) && (!filePath.isEmpty())) {
                File imageFile = new File(filePath);
                if (imageFile.exists() && imageFile.isFile()) {
                    InputStream inputStream = null;
                    try {
                        inputStream = new BufferedInputStream(new FileInputStream(imageFile));
                        byte[] buff = new byte[1024];
                        while ((inputStream.read(buff)) != -1) {
                            outputStream.write(buff);
                        }
                    } finally {
                        if (inputStream != null) {
                            inputStream.close ();
                        }
                    }
                }
            }
            log.liveImage = new SerialBlob(outputStream.toByteArray());
    
            logList.add(log);
        }

    } finally {
        reader.close();
    }
    return logList;
}

Read CSV

• L.67: Loading a CSV file instrument_log.csv with the data of the measurement log.
• L.70-74: read the contents of the CSV, it is set to an instance of InstrumentLog.
• L.77-95: read the file of the CSV of the live image file path, is writing the data to the Blob data.

List.4 TimeSeries Container Registration Process (InstrumentLogLogic.java)

public void registerTimeSeriesContainer(GridStore store, List<InstrumentLog> logList)
        throws GSException {
    for (InstrumentLog log : logList) {
        // Retrieve TimeSeries Container
        TimeSeries<InstrumentLog> logTs =
                store.putTimeSeries(log.weatherStationId, InstrumentLog.class);

        // Disable Auto Commit
        logTs.setAutoCommit(false);

        // Specify Time
        logTs.put(log.timestamp, log);

        // Specify Current Time
        // Comment out because the result would change on every execution
        // ts.append(log);

        // Instrument log timestamp is every 6 hours.
        // When the interpolation of 3 hours prior time to the log,
        // the temperature of the intermediate is registered
        Date medianTime = TimestampUtils.add(log.timestamp, -3, TimeUnit.HOUR);
        // Creating a TimeSeries Container when interpolating the values of temperature
        InstrumentLog medianLog = logTs.interpolate(medianTime, "temperture");
        // If there is no log, a NULL is returned before the interpolated time
        if (medianLog != null) {
            logTs.put(medianTime, medianLog);
        }

        // Commit
        logTs.commit();
    }
}

To register the data to the time series container in one of the following ways.

Specified time

• L.128: TimeSeries.put (Date, Class) method uses the time specified as the key when the data is registered.

Current time

• L.132: TimeSeries.append (Class) method uses the current time as the key. However, it has been commented out so that execution result is always constant.

Interpolated time

• L.137: TimestampUtils.add (Date, int, TimeUnit) specifies a time 3 hours prior to the time of the time of measurement log.
• L.139: TimeSeries.interpolate (Date, String) generates a linear interpolation of the value corresponding to the specified time. This means, for example, when the temperature is 70°F at 6 and 80°F at 12, the interpolated value would be 75°F if 9 was specified. To obtain a linear interpolation value must have adjacent keys or the same timestamp must have been previously registered.
• L.142: TimeSeries.put (Date, Class) and is registered with the data that is linear interpolation with method.

The following topics are described in more detail in the Collection Registration chapter:

• Auto-commit
• Commit
• Roll Back

Complete source code

Complete source code used in this sample can be downloaded from the following.

Download: timeseries-register.zip