## Key

• This line was removed.
• Formatting was changed.

This example illustrates how you can apply the following percentile-related functions to your transformations:

• `MEDIAN` - Calculate the median value from a column of values. See MEDIAN Function.
• `PERCENTILE` - Calculate a specified percentile for a column of values. See PERCENTILE Function.
• `QUARTILE` - Calculate a specified quartile for a column of values. See QUARTILE Function.

The following functions use an approximation technique for calculating median, percentile, and quartiles. In some cases, these calculations can be computed faster across large datasets.

• `APPROXIMATEMEDIAN` - Calculate a close approximation of the median value from a column of values. See APPROXIMATEMEDIAN Function.
• `APPROXIMATEPERCENTILE` - Calculate a close approximation of a specified percentile for a column of values. See APPROXIMATEPERCENTILE Function.
• `APPROXIMATEQUARTILE` - Calculate a close approximation of a specified quartile for a column of values. See APPROXIMATEQUARTILE Function.

Source:

The following table lists each student's height in inches:

StudentHeight
164
265
363
464
562
666
766
865
969
1066
1173
1269
1369
1461
1564
1661
1771
1867
1973
2066

Transformation:

Use the following transformations to calculate the median height in inches, a specified percentile and the first quartile.

• The first function uses a precise algorithm which can be slow to execute across large datasets.
• The second function uses an appropriate approximation algorithm, which is much faster to execute across large datasets.
• These approximate functions can use an error boundary parameter, which is set to `0.4` (0.4%) across all functions.

Median: This transformation calculates the median value, which corresponds to the 50th percentile.

D trans
RawWrangle true 'medianIn' step derive type: single value: median(heightIn) as: 'medianIn' Formula type Single row formula Formula median(heightIn) New column name New formula

D trans
RawWrangle true 'approxMedianIn' step derive type: single value: approxmedian(heightIn, 0.4) as: 'approxMedianIn' Formula type Single row formula Formula approximatemedian(heightIn, 0.4) New column name New formula

Percentile: This transformation calculates the 68th percentile.

D trans
RawWrangle true 'percentile68In' step derive type: single value: percentile(heightIn, 68, linear) as: 'percentile68In' Formula type Single row formula Formula percentile(heightIn, 68, linear) New column name New formula

D trans
RawWrangle true 'approxPercentile68In' step derive type: single value: approximatepercentile(heightIn, 68, 0.4) as: 'approxpercentile68In' Formula type Single row formula Formula approximatepercentile(heightIn, 68, 0.4) New column name New formula

Quartile: This transformation calculates the first quartile, which corresponds to the 25th percentile.

D trans
RawWrangle true 'percentile25In' step derive type: single value: quartile(heightIn, 1, linear) as: 'percentile25In' Formula type Single row formula Formula quartile(heightIn, 1, linear) New column name New formula

D trans
RawWrangle true 'approxPercentile25In' step derive type: single value: approximatequartile(heightIn, 1, 0.4) as: 'approxPercentile25In' Formula type Single row formula Formula approximatequartile(heightIn, 1, 0.4) New column name New formula

Results:

studentIdheightInapproxPercentile25Inpercentile25InapproxPercentile68Inpercentile68InapproxMedianInmedianIn
164646467.166.926666
265646467.166.926666
363646467.166.926666
464646467.166.926666
562646467.166.926666
666646467.166.926666
766646467.166.926666
865646467.166.926666
969646467.166.926666
1066646467.166.926666
1173646467.166.926666
1269646467.166.926666
1369646467.166.926666
1461646467.166.926666
1564646467.166.926666
1661646467.166.926666
1771646467.166.926666
1867646467.166.926666
1973646467.166.926666
2066646467.166.926666