Add float to string kernel

Signed-off-by: Haoyang Li <[email protected]>
NVIDIA · Oct 18, 2023 · cbce724 · cbce724
1 parent 6883988
commit cbce724
Show file tree

Hide file tree

Showing 3 changed files with 26 additions and 9 deletions.
diff --git a/src/main/cpp/src/cast_float_to_string.cu b/src/main/cpp/src/cast_float_to_string.cu
@@ -55,13 +55,13 @@ struct ftos_converter {
   static constexpr unsigned int significant_digits = 17;
   // maximum power-of-10 that will fit in 32-bits
   // static constexpr unsigned long long nine_digits = 1000000000;  // 1x10^9
-  static constexpr unsigned long long fifteen_digits = 1000000000000000;
+  // static constexpr unsigned long long fifteen_digits = 1000000000000000;
   static constexpr unsigned long long sixteen_digits = 10000000000000000;
   // Range of numbers here is for normalizing the value.
   // If the value is above or below the following limits, the output is converted to
   // scientific notation in order to show (at most) the number of significant digits.
-  static constexpr double upper_limit = 1000000000;  // max is 1x10^9
-  static constexpr double lower_limit = 0.0001;      // printf uses scientific notation below this
+  static constexpr double upper_limit = 10000000;  // max is 1x10^7
+  static constexpr double lower_limit = 0.001;      // printf uses scientific notation below this
   // Tables for doing normalization: converting to exponent form
   // IEEE double float has maximum exponent of 305 so these should cover everything
   double const upper10[9]  = {10, 100, 10000, 1e8, 1e16, 1e32, 1e64, 1e128, 1e256};
@@ -119,8 +119,16 @@ struct ftos_converter {
       }
     }
     //
-    int decimal_places = significant_digits - (exp10? 2 : 1);
-    unsigned long long max_digits = (exp10? fifteen_digits : sixteen_digits);
+    // int decimal_places = significant_digits - (exp10? 2 : 1);
+    // unsigned long long max_digits = (exp10? fifteen_digits : sixteen_digits);
+    int decimal_places = significant_digits - 1;
+    unsigned long long max_digits = sixteen_digits;
+    double temp_value = value;
+    while (temp_value < 1.0 && temp_value > 0.0) {
+      max_digits *= 10;
+      temp_value *= 10.0;
+      decimal_places++;
+    }
     integer                 = (unsigned int)value;
     for (unsigned int i = integer; i >= 10; i /= 10) {
       --decimal_places;
@@ -194,7 +202,7 @@ struct ftos_converter {
     // decimal
     *ptr++ = '.';
     if (decimal_places) {
-      char buffer[17];
+      char buffer[18];
       char* pb = buffer;
       while (decimal_places--) {
         *pb++ = (char)('0' + (decimal % 10));
@@ -232,7 +240,7 @@ struct ftos_converter {
       value = -value;
       bneg  = true;
     }
-    if (std::isinf(value)) return 3 + (int)bneg;  // Inf
+    if (std::isinf(value)) return 8 + (int)bneg;  // Inf
 
     // dissect float into parts
     unsigned int integer = 0;
@@ -261,7 +269,6 @@ struct ftos_converter {
         count ++;
         exp10 = -exp10;
       }
-      count += (int)(exp10 < 10);  // padding
       while (exp10 > 0) {
         exp10 /= 10;
         ++count;

diff --git a/src/main/java/com/nvidia/spark/rapids/jni/CastStrings.java b/src/main/java/com/nvidia/spark/rapids/jni/CastStrings.java
@@ -90,6 +90,16 @@ public static ColumnVector fromFloat(ColumnView cv) {
     return new ColumnVector(fromFloat(cv.getNativeView()));
   }
 
+  /**
+   * Convert a decimal column to a string column.
+   *
+   * @param cv the column data to process
+   * @return the converted column
+   */
+  public static ColumnVector fromDecimal(ColumnView cv) {
+    return new ColumnVector(fromDecimal(cv.getNativeView()));
+  }
+
   /**
    * Convert a string column to a given floating-point type column.
    *

diff --git a/thirdparty/cudf b/thirdparty/cudf
+5 −13		cpp/cmake/thirdparty/get_arrow.cmake
+2 −2		cpp/cmake/thirdparty/get_cufile.cmake
+2 −2		cpp/cmake/thirdparty/get_gtest.cmake
+5 −7		cpp/cmake/thirdparty/get_kvikio.cmake
+10 −8		cpp/cmake/thirdparty/get_libcudacxx.cmake
+1 −3		cpp/cmake/thirdparty/get_spdlog.cmake
+8 −7		cpp/cmake/thirdparty/get_thrust.cmake
+9 −31		cpp/include/cudf/detail/normalizing_iterator.cuh
+1 −5		cpp/include/cudf/lists/extract.hpp
+0 −6		cpp/include/cudf/lists/filling.hpp
+0 −2		cpp/include/cudf/lists/gather.hpp
+1 −3		cpp/include/cudf/lists/reverse.hpp
+1 −4		cpp/include/cudf/lists/sorting.hpp
+1 −5		cpp/include/cudf/lists/stream_compaction.hpp
+14 −18		cpp/include/cudf/strings/convert/convert_booleans.hpp
+14 −20		cpp/include/cudf/strings/convert/convert_datetime.hpp
+11 −15		cpp/include/cudf/strings/convert/convert_durations.hpp
+12 −18		cpp/include/cudf/strings/convert/convert_fixed_point.hpp
+12 −18		cpp/include/cudf/strings/convert/convert_floats.hpp
+31 −45		cpp/include/cudf/strings/convert/convert_integers.hpp
+12 −18		cpp/include/cudf/strings/convert/convert_ipv4.hpp
+6 −8		cpp/include/cudf/strings/convert/convert_lists.hpp
+9 −13		cpp/include/cudf/strings/convert/convert_urls.hpp
+0 −24		cpp/include/cudf/utilities/traits.hpp
+2 −2		cpp/src/lists/copying/segmented_gather.cu
+2 −4		cpp/src/lists/extract.cu
+2 −4		cpp/src/lists/reverse.cu
+3 −4		cpp/src/lists/segmented_sort.cu
+2 −4		cpp/src/lists/sequences.cu
+1 −2		cpp/src/lists/stream_compaction/apply_boolean_mask.cu
+1 −2		cpp/src/lists/stream_compaction/distinct.cu
+9 −11		cpp/src/strings/convert/convert_booleans.cu
+3 −6		cpp/src/strings/convert/convert_datetime.cu
+9 −11		cpp/src/strings/convert/convert_durations.cu
+4 −7		cpp/src/strings/convert/convert_fixed_point.cu
+19 −23		cpp/src/strings/convert/convert_floats.cu
+19 −17		cpp/src/strings/convert/convert_hex.cu
+56 −46		cpp/src/strings/convert/convert_integers.cu
+18 −21		cpp/src/strings/convert/convert_ipv4.cu
+1 −2		cpp/src/strings/convert/convert_lists.cu
+4 −6		cpp/src/strings/convert/convert_urls.cu
+1 −1		cpp/src/text/edit_distance.cu
+1 −14		cpp/src/utilities/traits.cpp
+0 −1		cpp/tests/CMakeLists.txt
+0 −1		cpp/tests/groupby/histogram_tests.cpp
+0 −100		cpp/tests/iterator/indexalator_test.cu
+0 −81		cpp/tests/streams/lists_test.cpp
+0 −146		cpp/tests/streams/strings/convert_test.cpp
+6 −27		cpp/tests/strings/booleans_tests.cpp
+3 −6		cpp/tests/strings/format_lists_tests.cpp
+0 −26		cpp/tests/strings/integers_tests.cpp
+2 −9		java/src/main/native/src/ColumnViewJni.cpp