<< Click to Display Table of Contents >> Navigation:  Reference > Commands > Value distribution tables

The form of the T command for a value distribution table is:

[+] T table [(F= ...)] [(ID='text')] = stub(max) * [banner] [+weight],

This is identical to a normal distribution table except that the stub must be and ivar, wvar or a cvar.

A distribution table is generated with the rows of the table being each separate value found in the data for the stub variable. At print time a row label is generated for each value found.

The maximum value (max) is the number of rows to be allocated for the table and must be large enough to hold each unique value found. If more values are found than rows allocated a warning is generated during execution and when printing the table.

If a cvar is used as the rows and no (max) setting is used then a default setting of (1000) will be used.

IMPORTANT: use F=RNA (Rank in Ascending Order) to reverse the row order if medians, quartiles, highest, lowest are used (formats MED, ILE, ILH or ILL).

You may wish to use F=NPSV (No Print Stats Value) to prevent the row values appearing twice.

You may wish to use F=NDIS (No DIStribution) to see only statistics based on the row values found.

IMPORTANT: if a cvar is used as the rows it should be exactly 20, 100 or 250 characters long. The RLW setting must be large enough to show all the text on one line; labels exceeding the RLW setting will be truncated without warning.

IMPORTANT: you must not use MT (manip table) commands on this type of table.

VDT tables can be overlaid; the same table number used more than once provided that the row type (and length if cvar) are the same and the same number of rows are allocated to each definition.

Examples of T value distribution tables:

t #freq=$ivol(2000)*$country,

t #median(f=med/ndis)=$ival(500)*$break,

 

dc $cuse(20),

dc $cuse=$ctitle,

t #title=$cuse(50)*,

 

dc $cusebig(250),

dc $cusebig=$long,

t #text=$cusebig(2000)*$break,