Signature | Description |
---|---|
enum class quantile_policy : unsigned char { lower_value = 1, // Take the higher index higher_value = 2, // Take the lower index mid_point = 3, // Average the two quantiles linear = 4, // Linearly combine the two quantiles }; |
This policy determines how to calculate quantiles when they fall between two values. Linear is calculates as: X1 + (X2 – X1) * (1.0 – QT) |
Signature | Description | Parameters |
---|---|---|
#include <DataFrame/DataFrameStatsVisitors.h> template<typename T, typename I = unsigned long> struct QuantileVisitor; // ------------------------------------- template<typename T, typename I = unsigned long> using qt_v = QuantileVisitor<T, I>; |
This is a “single action visitor”, meaning it is passed the whole data vector in one call and you must use the single_act_visit() interface. This functor class finds the quantile specified by quantile and q_policy. Please see quantile_policy for more explanation. QuantileVisitor (T quantile, quantile_policy q_policy); |
T: Column data type. I: Index type. |
static void test_quantile() { std::cout << "\nTesting QuantileVisitor{ } ..." << std::endl; std::vector<unsigned long> idx = { 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 31, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40 }; std::vector<double> d1 = { 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40 }; MyDataFrame df; df.load_data(std::move(idx), std::make_pair("col_1", d1)); df.shuffle<1, double>({"col_1"}, false); QuantileVisitor<double> v1(1, quantile_policy::mid_point); auto result = df.single_act_visit<double>("col_1", v1).get_result(); assert(result == 40.0); QuantileVisitor<double> v2(0.5, quantile_policy::mid_point); result = df.single_act_visit<double>("col_1", v2).get_result(); assert(result == 20.5); QuantileVisitor<double> v3(0.5, quantile_policy::linear); result = df.single_act_visit<double>("col_1", v3).get_result(); assert(result == 20.5); QuantileVisitor<double> v4(0.5, quantile_policy::higher_value); result = df.single_act_visit<double>("col_1", v4).get_result(); assert(result == 21.0); QuantileVisitor<double> v5(0.5, quantile_policy::lower_value); result = df.single_act_visit<double>("col_1", v5).get_result(); assert(result == 20.0); QuantileVisitor<double> v6(0.55, quantile_policy::mid_point); result = df.single_act_visit<double>("col_1", v6).get_result(); assert(result == 22.5); QuantileVisitor<double> v7(0.55, quantile_policy::linear); result = df.single_act_visit<double>("col_1", v7).get_result(); assert(result == 22.45); QuantileVisitor<double> v8(0.75, quantile_policy::mid_point); result = df.single_act_visit<double>("col_1", v8).get_result(); assert(result == 30.5); QuantileVisitor<double> v9(0.75, quantile_policy::linear); result = df.single_act_visit<double>("col_1", v9).get_result(); assert(result == 30.25); QuantileVisitor<double> v10(0, quantile_policy::linear); result = df.single_act_visit<double>("col_1", v10).get_result(); assert(result == 1.0); df.get_index().push_back(41); df.get_column<double>("col_1").push_back(41); QuantileVisitor<double> v11(0.75, quantile_policy::mid_point); result = df.single_act_visit<double>("col_1", v11).get_result(); assert(result == 31.0); QuantileVisitor<double> v12(0.75, quantile_policy::linear); result = df.single_act_visit<double>("col_1", v12).get_result(); assert(result == 31.0); QuantileVisitor<double> v13(0.75, quantile_policy::lower_value); result = df.single_act_visit<double>("col_1", v13).get_result(); assert(result == 31.0); QuantileVisitor<double> v14(0.75, quantile_policy::higher_value); result = df.single_act_visit<double>("col_1", v14).get_result(); assert(result == 31.0); QuantileVisitor<double> v15(0.71, quantile_policy::mid_point); result = df.single_act_visit<double>("col_1", v15).get_result(); assert(result == 29.5); QuantileVisitor<double> v16(0.71, quantile_policy::linear); result = df.single_act_visit<double>("col_1", v16).get_result(); assert(result == 29.29); QuantileVisitor<double> v17(0.23, quantile_policy::mid_point); result = df.single_act_visit<double>("col_1", v17).get_result(); assert(result == 9.5); QuantileVisitor<double> v18(0.2, quantile_policy::mid_point); result = df.single_act_visit<double>("col_1", v18).get_result(); assert(result == 8.5); QuantileVisitor<double> v19(0.23, quantile_policy::linear); result = df.single_act_visit<double>("col_1", v19).get_result(); assert(result == 9.77); QuantileVisitor<double> v20(0.23, quantile_policy::lower_value); result = df.single_act_visit<double>("col_1", v20).get_result(); assert(result == 9.0); QuantileVisitor<double> v21(0.23, quantile_policy::higher_value); result = df.single_act_visit<double>("col_1", v21).get_result(); assert(result == 10.0); QuantileVisitor<double> v22(1, quantile_policy::linear); result = df.single_act_visit<double>("col_1", v22).get_result(); assert(result == 41.0); QuantileVisitor<double> v23(0, quantile_policy::mid_point); result = df.single_act_visit<double>("col_1", v23).get_result(); assert(result == 1.0); }