Remove INT96 timestamps in cuDF Parquet writer

cpp/include/cudf/io/parquet.hpp

@@ -596,9 +596,6 @@ class parquet_writer_options {
   std::optional<table_input_metadata> _metadata;
   // Optional footer key_value_metadata
   std::vector<std::map<std::string, std::string>> _user_data;
-  // Parquet writer can write INT96 or TIMESTAMP_MICROS. Defaults to TIMESTAMP_MICROS.
-  // If true then overrides any per-column setting in _metadata.
-  bool _write_timestamps_as_int96 = false;
   // Parquet writer can write timestamps as UTC
   // Defaults to true because libcudf timestamps are implicitly UTC
   bool _write_timestamps_as_UTC = true;
@@ -717,13 +714,6 @@ class parquet_writer_options {
     return _user_data;
   }
 
-  /**
-   * @brief Returns `true` if timestamps will be written as INT96
-   *
-   * @return `true` if timestamps will be written as INT96
-   */
-  bool is_enabled_int96_timestamps() const { return _write_timestamps_as_int96; }
-
   /**
    * @brief Returns `true` if timestamps will be written as UTC
    *
@@ -867,14 +857,6 @@ class parquet_writer_options {
    */
   void set_compression(compression_type compression) { _compression = compression; }
 
-  /**
-   * @brief Sets timestamp writing preferences. INT96 timestamps will be written
-   * if `true` and TIMESTAMP_MICROS will be written if `false`.
-   *
-   * @param req Boolean value to enable/disable writing of INT96 timestamps
-   */
-  void enable_int96_timestamps(bool req) { _write_timestamps_as_int96 = req; }
-
   /**
    * @brief Sets preference for writing timestamps as UTC. Write timestamps as UTC if set to `true`.
    *
@@ -1191,18 +1173,6 @@ class parquet_writer_options_builder {
     return *this;
   }
 
-  /**
-   * @brief Sets whether int96 timestamps are written or not in parquet_writer_options.
-   *
-   * @param enabled Boolean value to enable/disable int96 timestamps
-   * @return this for chaining
-   */
-  parquet_writer_options_builder& int96_timestamps(bool enabled)
-  {
-    options._write_timestamps_as_int96 = enabled;
-    return *this;
-  }
-
   /**
    * @brief Set to true if timestamps are to be written as UTC.
    *
@@ -1293,9 +1263,6 @@ class chunked_parquet_writer_options {
   std::optional<table_input_metadata> _metadata;
   // Optional footer key_value_metadata
   std::vector<std::map<std::string, std::string>> _user_data;
-  // Parquet writer can write INT96 or TIMESTAMP_MICROS. Defaults to TIMESTAMP_MICROS.
-  // If true then overrides any per-column setting in _metadata.
-  bool _write_timestamps_as_int96 = false;
   // Parquet writer can write timestamps as UTC. Defaults to true.
   bool _write_timestamps_as_UTC = true;
   // Maximum size of each row group (unless smaller than a single page)
@@ -1376,13 +1343,6 @@ class chunked_parquet_writer_options {
     return _user_data;
   }
 
-  /**
-   * @brief Returns `true` if timestamps will be written as INT96
-   *
-   * @return `true` if timestamps will be written as INT96
-   */
-  bool is_enabled_int96_timestamps() const { return _write_timestamps_as_int96; }
-
   /**
    * @brief Returns `true` if timestamps will be written as UTC
    *
@@ -1509,15 +1469,6 @@ class chunked_parquet_writer_options {
    */
   void set_compression(compression_type compression) { _compression = compression; }
 
-  /**
-   * @brief Sets timestamp writing preferences.
-   *
-   * INT96 timestamps will be written if `true` and TIMESTAMP_MICROS will be written if `false`.
-   *
-   * @param req Boolean value to enable/disable writing of INT96 timestamps
-   */
-  void enable_int96_timestamps(bool req) { _write_timestamps_as_int96 = req; }
-
   /**
    * @brief Sets preference for writing timestamps as UTC. Write timestamps as UTC if set to `true`.
    *
@@ -1684,21 +1635,6 @@ class chunked_parquet_writer_options_builder {
     return *this;
   }
 
-  /**
-   * @brief Set to true if timestamps should be written as
-   * int96 types instead of int64 types. Even though int96 is deprecated and is
-   * not an internal type for cudf, it needs to be written for backwards
-   * compatibility reasons.
-   *
-   * @param enabled Boolean value to enable/disable int96 timestamps
-   * @return this for chaining
-   */
-  chunked_parquet_writer_options_builder& int96_timestamps(bool enabled)
-  {
-    options._write_timestamps_as_int96 = enabled;
-    return *this;
-  }
-
   /**
    * @brief Set to true if timestamps are to be written as UTC.
    *

cpp/include/cudf/io/types.hpp

@@ -605,10 +605,9 @@ class column_in_metadata {
   friend table_input_metadata;
   std::string _name = "";
   std::optional<bool> _nullable;
-  bool _list_column_is_map  = false;
-  bool _use_int96_timestamp = false;
-  bool _output_as_binary    = false;
-  bool _skip_compression    = false;
+  bool _list_column_is_map = false;
+  bool _output_as_binary   = false;
+  bool _skip_compression   = false;
   std::optional<uint8_t> _decimal_precision;
   std::optional<int32_t> _parquet_field_id;
   std::optional<int32_t> _type_length;
@@ -672,20 +671,6 @@ class column_in_metadata {
     return *this;
   }
 
-  /**
-   * @brief Specifies whether this timestamp column should be encoded using the deprecated int96
-   * physical type. Only valid for the following column types:
-   * timestamp_s, timestamp_ms, timestamp_us, timestamp_ns
-   *
-   * @param req True = use int96 physical type. False = use int64 physical type
-   * @return this for chaining
-   */
-  column_in_metadata& set_int96_timestamps(bool req) noexcept
-  {
-    _use_int96_timestamp = req;
-    return *this;
-  }
-
   /**
    * @brief Set the decimal precision of this column. Only valid if this column is a decimal
    * (fixed-point) type
@@ -818,14 +803,6 @@ class column_in_metadata {
    */
   [[nodiscard]] bool is_map() const noexcept { return _list_column_is_map; }
 
-  /**
-   * @brief Get whether to encode this timestamp column using deprecated int96 physical type
-   *
-   * @return Boolean indicating whether to encode this timestamp column using deprecated int96
-   *         physical type
-   */
-  [[nodiscard]] bool is_enabled_int96_timestamps() const noexcept { return _use_int96_timestamp; }
-
   /**
    * @brief Get whether precision has been set for this decimal column
    *

cpp/src/io/parquet/chunk_dict.cu

@@ -145,7 +145,6 @@ CUDF_KERNEL void __launch_bounds__(block_size)
         switch (col->physical_type) {
           case Type::INT32: return 4;
           case Type::INT64: return 8;
-          case Type::INT96: return 12;
           case Type::FLOAT: return 4;
           case Type::DOUBLE: return 8;
           case Type::BYTE_ARRAY: {

cpp/src/io/parquet/page_enc.cu

@@ -115,7 +115,6 @@ constexpr uint32_t physical_type_len(Type physical_type, type_id id, int type_le
     return id == type_id::DECIMAL128 ? sizeof(__int128_t) : type_length;
   }
   switch (physical_type) {
-    case INT96: return 12u;
     case INT64:
     case DOUBLE: return sizeof(int64_t);
     case BOOLEAN: return 1u;
@@ -545,7 +544,6 @@ __device__ size_t delta_data_len(Type physical_type,
   auto const dtype_len_out = physical_type_len(physical_type, type_id, sizeof(int32_t));
   auto const dtype_len     = [&]() -> uint32_t {
     if (physical_type == INT32) { return int32_logical_len(type_id); }
-    if (physical_type == INT96) { return sizeof(int64_t); }
     return dtype_len_out;
   }();
 
@@ -1324,27 +1322,6 @@ constexpr auto julian_calendar_epoch_diff()
   return sys_days{January / 1 / 1970} - (sys_days{November / 24 / -4713} + 12h);
 }
 
-/**
- * @brief Converts number `v` of periods of type `PeriodT` into a pair with nanoseconds since
- * midnight and number of Julian days. Does not deal with time zones. Used by INT96 code.
- *
- * @tparam PeriodT a ratio representing the tick period in duration
- * @param v count of ticks since epoch
- * @return A pair of (nanoseconds, days) where nanoseconds is the number of nanoseconds
- * elapsed in the day and days is the number of days from Julian epoch.
- */
-template <typename PeriodT>
-__device__ auto julian_days_with_time(int64_t v)
-{
-  using namespace cuda::std::chrono;
-  auto const dur_total             = duration<int64_t, PeriodT>{v};
-  auto const dur_days              = floor<days>(dur_total);
-  auto const dur_time_of_day       = dur_total - dur_days;
-  auto const dur_time_of_day_nanos = duration_cast<nanoseconds>(dur_time_of_day);
-  auto const julian_days           = dur_days + ceil<days>(julian_calendar_epoch_diff());
-  return std::make_pair(dur_time_of_day_nanos, julian_days);
-}
-
 // this has been split out into its own kernel because of the amount of shared memory required
 // for the state buffer. encode kernels that don't use the RLE buffer can get started while
 // the level data is encoded.
@@ -1666,7 +1643,6 @@ CUDF_KERNEL void __launch_bounds__(block_size, 8)
   auto const dtype_len_out = physical_type_len(physical_type, type_id, s->col.type_length);
   auto const dtype_len_in  = [&]() -> uint32_t {
     if (physical_type == INT32) { return int32_logical_len(type_id); }
-    if (physical_type == INT96) { return sizeof(int64_t); }
     return dtype_len_out;
   }();
 
@@ -1770,40 +1746,6 @@ CUDF_KERNEL void __launch_bounds__(block_size, 8)
           }
           encode_value(dst + pos, v, stride);
         } break;
-        case INT96: {
-          // only PLAIN encoding is supported
-          int64_t v        = s->col.leaf_column->element<int64_t>(val_idx);
-          int32_t ts_scale = s->col.ts_scale;
-          if (ts_scale != 0) {
-            if (ts_scale < 0) {
-              v /= -ts_scale;
-            } else {
-              v *= ts_scale;
-            }
-          }
-
-          auto const [last_day_nanos, julian_days] = [&] {
-            using namespace cuda::std::chrono;
-            switch (s->col.leaf_column->type().id()) {
-              case type_id::TIMESTAMP_SECONDS:
-              case type_id::TIMESTAMP_MILLISECONDS: {
-                return julian_days_with_time<cuda::std::milli>(v);
-              } break;
-              case type_id::TIMESTAMP_MICROSECONDS:
-              case type_id::TIMESTAMP_NANOSECONDS: {
-                return julian_days_with_time<cuda::std::micro>(v);
-              } break;
-            }
-            return julian_days_with_time<cuda::std::nano>(0);
-          }();
-
-          // the 12 bytes of fixed length data.
-          v = last_day_nanos.count();
-          encode_value(dst + pos, v, 1);
-          uint32_t w = julian_days.count();
-          encode_value(dst + pos + 8, w, 1);
-        } break;
-
         case BYTE_ARRAY: {
           // only PLAIN encoding is supported
           auto const bytes = [](cudf::type_id const type_id,
@@ -1901,7 +1843,6 @@ CUDF_KERNEL void __launch_bounds__(block_size, 8)
   auto const dtype_len_out = physical_type_len(physical_type, type_id, s->col.type_length);
   auto const dtype_len_in  = [&]() -> uint32_t {
     if (physical_type == INT32) { return int32_logical_len(type_id); }
-    if (physical_type == INT96) { return sizeof(int64_t); }
     return dtype_len_out;
   }();
 
@@ -2033,7 +1974,6 @@ CUDF_KERNEL void __launch_bounds__(block_size, 8)
   auto const dtype_len_out = physical_type_len(physical_type, type_id, s->col.type_length);
   auto const dtype_len_in  = [&]() -> uint32_t {
     if (physical_type == INT32) { return int32_logical_len(type_id); }
-    if (physical_type == INT96) { return sizeof(int64_t); }
     return dtype_len_out;
   }();
 

cpp/src/io/parquet/predicate_pushdown.cpp

@@ -86,10 +86,6 @@ struct stats_caster {
     switch (type) {
       case INT32: return targetType<T>(*reinterpret_cast<int32_t const*>(stats_val));
       case INT64: return targetType<T>(*reinterpret_cast<int64_t const*>(stats_val));
-      case INT96:  // Deprecated in parquet specification
-        return targetType<T>(static_cast<__int128_t>(reinterpret_cast<int64_t const*>(stats_val)[0])
-                               << 32 |
-                             reinterpret_cast<int32_t const*>(stats_val)[2]);
       case BYTE_ARRAY: [[fallthrough]];
       case FIXED_LEN_BYTE_ARRAY:
         if (stats_size == sizeof(T)) {