Math Functions > Time Functions


Extract Time Series

The Extract Time Series function selects/extracts data points from a regular or irregular interval time series data set based on user defined time specifications.  For example, the function may be used to extract values observed every day at noon from hourly interval data.

 

Data may be extracted on the basis of one of the following time specifications: year, month of the year (January, February, etc.), day of month, day of week (Sunday, Monday, etc.), or time of day.

 

The appearance of the Math Functions screen for the Extract Time Series function is shown below. 

 

Screen for Extract Time Series Function

 

To derive a new time series data set from the selected times of another time series data set:

  1. Choose the Time Functions tab of the Math Functions screen and select the Extract Time Series operator.
  2. From the Selected Data Set list, select a time series data set to apply the function. 
  3. Choose one of the time levels, Year, Month, Day of Month, Day of Week, or Time of day and complete the boxes to the right.  Entering or selecting a value in only the left-most box sets a time specification for data extraction to an individual year, month, day, or time of day.  Fill in both boxes to specify a time range for data extraction.
  4. Click Year to extract data for an individual year or range of years.  To extract data for an individual year, fill the left-most box in with the desired year.  To extract data for a range of years fill the left box with the beginning year and the right box with the ending year.
  5. Click Month to extract data for an individual month of the year or range of months.  Use the lists to set the month boxes.  The list contains the months of the year, JAN to DEC.  Set only the left-most box to extract data for a single month of the year.  Set both boxes to extract data for a range of months.  The range of months can be set to encompass the end of the year; that is, “OCT to FEB” is a valid setting.
  6. Click Day of Month to extract data for an individual day or range of day of month days.  The days are set using the two lists to the right.  As above, the lists can be used to specify a single day of month, using the left-most pull down, or a range of days, using both lists.  The day of month values range from 1 to 31, and “LastDay,” the last day of the month, which may vary by month.  The range of days can be set to encompass the end of the month.  For example, for “27 to 4,” data will be extracted from the 27th of one month to the 4th of the next month.
  7. Click Day of Week to extract data for an individual or range of day of the weekdays.  Select day from the two lists to the right.  As above, you may use the lists to specify one day, using the left-most list, or a range of days, using both lists.  The day of week values range from SUN to SAT.  The range of days can be set to encompass the end of the week; that is, “SAT to MON” is a valid setting.
  8. Click Time to extract data by the time of day.  An individual time is specified by completing the left-most box only.  Complete both boxes to specify a time of day range for data extraction.  The time specification employs the standard four-digit military style 24-hour format (e.g. “2300” or “0310 to 0600”).  The time range can encompass the end of day; that is, “2200 to 0330” is a valid setting.  The optional Time Window box is applied to the time of day extraction only.  The Time Window is an integer value in minutes, used to extend the beginning and ending of the time of day period for data extraction.  For example, with a time of day extraction time of “0300” and a Time Window of “10 minutes,” data will be extracted from the selected time series if times falls within in the period 0250 to 0310.
  9. Select the Set as Irregular checkbox (recommended) to ensure the resultant time series data set is identified as irregular interval time series data.  If the box is not checked, the function will attempt to determine if the extracted data set can be classified as regular interval time series data.