Remove cudf::strings::repeat_strings_output_sizes and optional parameter from cudf::strings::repeat_strings

cpp/benchmarks/string/repeat_strings.cpp

@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 2021-2022, NVIDIA CORPORATION.
+ * Copyright (c) 2021-2023, NVIDIA CORPORATION.
  *
  * Licensed under the Apache License, Version 2.0 (the "License");
  * you may not use this file except in compliance with the License.
@@ -79,42 +79,6 @@ static void BM_repeat_strings_column_times(benchmark::State& state)
                           (strings_col.chars_size() + repeat_times_col.size() * sizeof(int32_t)));
 }
 
-static void BM_compute_output_strings_sizes(benchmark::State& state)
-{
-  auto const n_rows           = static_cast<cudf::size_type>(state.range(0));
-  auto const max_str_length   = static_cast<cudf::size_type>(state.range(1));
-  auto const table            = create_data_table(2, n_rows, max_str_length);
-  auto const strings_col      = cudf::strings_column_view(table->view().column(0));
-  auto const repeat_times_col = table->view().column(1);
-
-  for ([[maybe_unused]] auto _ : state) {
-    [[maybe_unused]] cuda_event_timer raii(state, true, cudf::get_default_stream());
-    cudf::strings::repeat_strings_output_sizes(strings_col, repeat_times_col);
-  }
-
-  state.SetBytesProcessed(state.iterations() *
-                          (strings_col.chars_size() + repeat_times_col.size() * sizeof(int32_t)));
-}
-
-static void BM_repeat_strings_column_times_precomputed_sizes(benchmark::State& state)
-{
-  auto const n_rows           = static_cast<cudf::size_type>(state.range(0));
-  auto const max_str_length   = static_cast<cudf::size_type>(state.range(1));
-  auto const table            = create_data_table(2, n_rows, max_str_length);
-  auto const strings_col      = cudf::strings_column_view(table->view().column(0));
-  auto const repeat_times_col = table->view().column(1);
-  [[maybe_unused]] auto const [sizes, total_bytes] =
-    cudf::strings::repeat_strings_output_sizes(strings_col, repeat_times_col);
-
-  for ([[maybe_unused]] auto _ : state) {
-    [[maybe_unused]] cuda_event_timer raii(state, true, cudf::get_default_stream());
-    cudf::strings::repeat_strings(strings_col, repeat_times_col, *sizes);
-  }
-
-  state.SetBytesProcessed(state.iterations() *
-                          (strings_col.chars_size() + repeat_times_col.size() * sizeof(int32_t)));
-}
-
 static void generate_bench_args(benchmark::internal::Benchmark* b)
 {
   int const min_rows   = 1 << 8;
@@ -145,23 +109,5 @@ class RepeatStrings : public cudf::benchmark {
     ->UseManualTime()                                               \
     ->Unit(benchmark::kMillisecond);
 
-#define COMPUTE_OUTPUT_STRINGS_SIZES_BENCHMARK_DEFINE(name)          \
-  BENCHMARK_DEFINE_F(RepeatStrings, name)                            \
-  (::benchmark::State & st) { BM_compute_output_strings_sizes(st); } \
-  BENCHMARK_REGISTER_F(RepeatStrings, name)                          \
-    ->Apply(generate_bench_args)                                     \
-    ->UseManualTime()                                                \
-    ->Unit(benchmark::kMillisecond);
-
-#define REPEAT_STRINGS_COLUMN_TIMES_PRECOMPUTED_SIZES_BENCHMARK_DEFINE(name)          \
-  BENCHMARK_DEFINE_F(RepeatStrings, name)                                             \
-  (::benchmark::State & st) { BM_repeat_strings_column_times_precomputed_sizes(st); } \
-  BENCHMARK_REGISTER_F(RepeatStrings, name)                                           \
-    ->Apply(generate_bench_args)                                                      \
-    ->UseManualTime()                                                                 \
-    ->Unit(benchmark::kMillisecond);
-
 REPEAT_STRINGS_SCALAR_TIMES_BENCHMARK_DEFINE(scalar_times)
 REPEAT_STRINGS_COLUMN_TIMES_BENCHMARK_DEFINE(column_times)
-COMPUTE_OUTPUT_STRINGS_SIZES_BENCHMARK_DEFINE(compute_output_strings_sizes)
-REPEAT_STRINGS_COLUMN_TIMES_PRECOMPUTED_SIZES_BENCHMARK_DEFINE(precomputed_sizes)

cpp/include/cudf/strings/repeat_strings.hpp

@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 2021-2022, NVIDIA CORPORATION.
+ * Copyright (c) 2021-2023, NVIDIA CORPORATION.
  *
  * Licensed under the Apache License, Version 2.0 (the "License");
  * you may not use this file except in compliance with the License.
@@ -32,15 +32,15 @@ namespace strings {
  */
 
 /**
- * @brief Repeat the given string scalar by a given number of times.
+ * @brief Repeat the given string scalar a given number of times
  *
  * An output string scalar is generated by repeating the input string by a number of times given by
- * the @p `repeat_times` parameter.
+ * the `repeat_times` parameter.
  *
  * In special cases:
- *  - If @p `repeat_times` is not a positive value, an empty (valid) string scalar will be returned.
+ *  - If `repeat_times` is not a positive value, an empty (valid) string scalar will be returned.
  *  - An invalid input scalar will always result in an invalid output scalar regardless of the
- *    value of @p `repeat_times` parameter.
+ *    value of `repeat_times` parameter.
  *
  * @code{.pseudo}
  * Example:
@@ -50,33 +50,30 @@ namespace strings {
  * @endcode
  *
  * @throw cudf::logic_error if the size of the output string scalar exceeds the maximum value that
- *        can be stored by the index type
- *        (i.e., @code input.size() * repeat_times > numeric_limits<size_type>::max() @endcode).
+ *        can be stored by the index type:
+ *        `input.size() * repeat_times > max of size_type`
  *
- * @param input The scalar containing the string to repeat.
- * @param repeat_times The number of times the input string is repeated.
- * @param mr Device memory resource used to allocate the returned string scalar.
- * @return New string scalar in which the input string is repeated.
+ * @param input The scalar containing the string to repeat
+ * @param repeat_times The number of times the input string is repeated
+ * @param mr Device memory resource used to allocate the returned string scalar
+ * @return New string scalar in which the input string is repeated
  */
 std::unique_ptr<string_scalar> repeat_string(
   string_scalar const& input,
   size_type repeat_times,
   rmm::mr::device_memory_resource* mr = rmm::mr::get_current_device_resource());
 
 /**
- * @brief Repeat each string in the given strings column by a given number of times.
+ * @brief Repeat each string in the given strings column a given number of times
  *
- * An output strings column is generated by repeating each string from the input strings column by a
- * number of times given by the @p `repeat_times` parameter.
+ * An output strings column is generated by repeating each string from the input strings column by
+ * the number of times given by the `repeat_times` parameter.
  *
  * In special cases:
- *  - If @p `repeat_times` is not a positive number, a non-null input string will always result in
+ *  - If `repeat_times` is not a positive number, a non-null input string will always result in
  *    an empty output string.
  *  - A null input string will always result in a null output string regardless of the value of the
- *    @p `repeat_times` parameter.
- *
- * The caller is responsible for checking the output column size will not exceed the maximum size of
- * a strings column (number of total characters is less than the max size_type value).
+ *    `repeat_times` parameter.
  *
  * @code{.pseudo}
  * Example:
@@ -85,10 +82,10 @@ std::unique_ptr<string_scalar> repeat_string(
  * out is ['aaaaaa', null, '', 'bbcbbcbbc']
  * @endcode
  *
- * @param input The column containing strings to repeat.
- * @param repeat_times The number of times each input string is repeated.
- * @param mr Device memory resource used to allocate the returned strings column.
- * @return New column containing the repeated strings.
+ * @param input The column containing strings to repeat
+ * @param repeat_times The number of times each input string is repeated
+ * @param mr Device memory resource used to allocate the returned strings column
+ * @return New column containing the repeated strings
  */
 std::unique_ptr<column> repeat_strings(
   strings_column_view const& input,
@@ -97,21 +94,17 @@ std::unique_ptr<column> repeat_strings(
 
 /**
  * @brief Repeat each string in the given strings column by the numbers of times given in another
- * numeric column.
+ * numeric column
  *
  * An output strings column is generated by repeating each of the input string by a number of times
- * given by the corresponding row in a @p `repeat_times` numeric column. The computational time can
- * be reduced if sizes of the output strings are known and provided.
+ * given by the corresponding row in a `repeat_times` numeric column.
  *
  * In special cases:
  *  - Any null row (from either the input strings column or the `repeat_times` column) will always
  *    result in a null output string.
  *  - If any value in the `repeat_times` column is not a positive number and its corresponding input
  *    string is not null, the output string will be an empty string.
  *
- * The caller is responsible for checking the output column size will not exceed the maximum size of
- * a strings column (number of total characters is less than the max size_type value).
- *
  * @code{.pseudo}
  * Example:
  * strs         = ['aa', null, '', 'bbc-']
@@ -120,51 +113,16 @@ std::unique_ptr<column> repeat_strings(
  * out is ['aa', null, '', 'bbc-bbc-bbc-bbc-']
  * @endcode
  *
- * @throw cudf::logic_error if the input `repeat_times` column has data type other than integer.
+ * @throw cudf::logic_error if the input `repeat_times` is not an integer type
  * @throw cudf::logic_error if the input columns have different sizes.
  *
- * @param input The column containing strings to repeat.
+ * @param input The column containing strings to repeat
  * @param repeat_times The column containing numbers of times that the corresponding input strings
- *        are repeated.
- * @param output_strings_sizes The optional column containing pre-computed sizes of the output
- *        strings.
- * @param mr Device memory resource used to allocate the returned strings column.
+ *                     are repeated
+ * @param mr Device memory resource used to allocate the returned strings column
  * @return New column containing the repeated strings.
  */
 std::unique_ptr<column> repeat_strings(
-  strings_column_view const& input,
-  column_view const& repeat_times,
-  std::optional<column_view> output_strings_sizes = std::nullopt,
-  rmm::mr::device_memory_resource* mr             = rmm::mr::get_current_device_resource());
-
-/**
- * @brief Compute sizes of the output strings if each string in the input strings column
- * is repeated by the numbers of times given in another numeric column.
- *
- * The output column storing string output sizes is not nullable. These string sizes are
- * also summed up and returned (in an `int64_t` value), which can be used to detect if the input
- * strings column can be safely repeated without data corruption due to overflow in string indexing.
- *
- * @code{.pseudo}
- * Example:
- * strs         = ['aa', null, '', 'bbc-']
- * repeat_times = [ 1,   2,     3,  4   ]
- * [output_sizes, total_size] = repeat_strings_output_sizes(strs, repeat_times)
- * out is [2, 0, 0, 16], and total_size = 18
- * @endcode
- *
- * @throw cudf::logic_error if the input `repeat_times` column has data type other than integer.
- * @throw cudf::logic_error if the input columns have different sizes.
- *
- * @param input The column containing strings to repeat.
- * @param repeat_times The column containing numbers of times that the corresponding input strings
- *        are repeated.
- * @param mr Device memory resource used to allocate the returned strings column.
- * @return A pair with the first item is an int32_t column containing sizes of the output strings,
- *         and the second item is an int64_t number containing the total sizes (in bytes) of the
- *         output strings column.
- */
-std::pair<std::unique_ptr<column>, int64_t> repeat_strings_output_sizes(
   strings_column_view const& input,
   column_view const& repeat_times,
   rmm::mr::device_memory_resource* mr = rmm::mr::get_current_device_resource());