Skip to content

TSTool / Command / SplitTableRow


The SplitTableRow command uses information from a table row to create a sequence of new table rows. The new rows are inserted after the original row, which optionally can be deleted from the table after processing. The table rows can be split using either “tuple” or “measure” approach as described below:

Tuple Split

A “tuple” is a sequence of related values and contain 1 or more grouped values. For example, the following figure illustrates annual time series values listed horizontally in each row.

SplitTableRow Tuple Input

This command can split the data into the following representation, which is more conducive to time series processing (for example use the TableToTimeSeries command to convert the results to time series that can be processed in TSTool).

SplitTableRow Tuple Output

Measure Split

A measure split is used to split a longer distance into shorter segments. Start and end values in the row are used to create segmented row distances. For example, this can be used to break a streamflow reach into segments of equal length, for example “stream mile” segments as shown in the following example. The first table illustrates input and the second the table after modification (original rows are deleted). The start and end columns can be specified in any order but the output is always with the smaller value as the start.

SplitTableRow Split Input

SplitTableRow Split Output

Command Editor

The following dialog is used to edit the command and illustrates the syntax of the command for tuple split.

SplitTableRow Tuple

SplitTableRow Command Editor for Tuple Split (see also the full-size image)

The following dialog is used to edit the command and illustrates the syntax of the command for measure split.


SplitTableRow Command Editor for Measure Split (see also the full-size image)

Command Syntax

The command syntax is as follows:


Command Parameters

Method Parameter                                     Description Default          
All TableID
The identifier for the table. None – must be specified.
All DeleteOriginalRow Indicate whether the original table row should be deleted as False or True. Specify True if the generated sequence of rows should replace the original row. False
Tuple TupleColumns The names of columns in each tuple, where the tuple groups are separated by semi-colons and the columns within the tuple are separated by commas.
For a 1-tuple: Column1;Column;Column3
For a 2-tuple: Column1a,Column1b;Column2a,Column2b
None – must be specified for tuple approach.
Tuple TupleDateTimes The date/times corresponding to each tuple, which will be matched with values inserted into the NewTupleDateTimeColumn. No date/time is associated with tuples.
Tuple NewTupleColumns The output columns to be used when splitting the tuples.
For a 1-tuple: NewColumn1
For a 2-tuple: NewColumn1,NewColumn2
None – must be specified for tuple approach.
Tuple NewTupleDateTimeColumn If TupleDateTimes is specified, this column will be filled with a data/time corresponding to each input tuple. No date/time output.
Tuple InsertBeforeColumn Indicate the name of the column before which to insert new columns. Currently this is used only with tuples. Append at end.
Measure MeasureStartColumn The name of the table column for the starting measure. None – must be specified.
Measure MeasureEndColumn The name of the table column for the ending measure. None – must be specified.
Measure MeasureIncrement The measure increment used to split the original distance into segments (e.g., .1). None – must be specified.
Measure MinimumStartSegmentLength The minimum length of the starting segment to include. Include start segment.
Measure MinimumEndSegmentLength The minimum length of the ending segment to include. Include end segment.


See the automated tests.


See Also