TD_UnivariateStatistics Syntax Elements - Teradata® Database

TargetColumns

Specify the names of the numeric InputTable columns for which to compute statistics.

PartitionColumns

[Optional] Specify the names of the InputTable columns on which to partition the input. The function copies these columns to the output table.

Default behavior: The function treats all rows as a single partition.

Stats

[Optional] Specify the statistics to calculate. statistic is one of these:

• SUM
• COUNT or CNT
• MAXIMUM or MAX
• MINIMUM or MIN
• MEAN
• UNCORRECTED SUM OF SQUARES or USS
• NULL COUNT or NLC
• POSITIVE VALUES COUNT or PVC
• NEGATIVE VALUES COUNT or NVC
• ZERO VALUES COUNT or ZVC
• TOP5 or TOP
• BOTTOM5 or BTM
• RANGE or RNG
• GEOMETRIC MEAN or GM
• HARMONIC MEAN or HM
• VARIANCE or VAR
• STANDARD DEVIATION or STD
• STANDARD ERROR or SE
• SKEWNESS or SKW
• KURTOSIS or KUR
• COEFFICIENT OF VARIATION or CV
• CORRECTED SUM OF SQUARES or CSS
• MODE
• MEDIAN or MED
• UNIQUE ENTITY COUNT or UEC
• INTERQUARTILE RANGE or IQR
• TRIMMED MEAN or TM
• PERCENTILES or PRC
• ALL

Default: ALL

Centiles

[Optional] Specify the centile to calculate. percentile is an INTEGER in the range [1, 100].

The function ignores Centiles unless Stats specifies PERCENTILES, PRC, or ALL.

Default: 1, 5, 10, 25, 50, 75, 90, 95, 99

TrimPercentile

[Optional] Specify the trimmed lower percentile, an integer value in the range [1, 50]. The function calculates the mean of the values between the trimmed lower percentile (trimmed_percentile) and trimmed upper percentile (1-trimmed_percentile).

The function ignores TrimPercentile unless Stats specifies TRIMMED MEAN, TM, or ALL.

Default: 20