[Time Conversion] Timestamp to time

Devjiu · Devjiu · commit d14f20866bdb · 2023-06-09T09:03:43.000-05:00
Adding timestamp to time conversion. The units is also taken into account. Resolves: #318 Signed-off-by: Dmitrii Makarenko <dmitrii.makarenko@intel.com>
diff --git a/omniscidb/IR/Expr.cpp b/omniscidb/IR/Expr.cpp
@@ -217,7 +217,9 @@ bool isCastAllowed(const Type* old_type, const Type* new_type) {
     // can cast from date to timestamp
   } else if (old_type->isDate() && new_type->isTimestamp()) {
     return true;
-  } else if (old_type->isTimestamp() && new_type->isDate()) {
+  } else if (old_type->isTimestamp() && (new_type->isDate() || new_type->isTime())) {
+    return true;
+  } else if (old_type->isTime() && new_type->isInteger()) {
     return true;
   } else if (old_type->isBoolean() && new_type->isNumber()) {
     return true;
diff --git a/omniscidb/QueryBuilder/QueryBuilder.cpp b/omniscidb/QueryBuilder/QueryBuilder.cpp
@@ -915,11 +915,11 @@ BuilderExpr BuilderExpr::cast(const Type* new_type) const {
       return {builder_, expr_->cast(new_type), "", true};
     }
   } else if (expr_->type()->isTime()) {
-    if (new_type->isTime()) {
+    if (new_type->isInteger() || new_type->isTime()) {
       return {builder_, expr_->cast(new_type), "", true};
     }
   } else if (expr_->type()->isTimestamp()) {
-    if (new_type->isNumber() || new_type->isDate() || new_type->isTimestamp()) {
+    if (new_type->isInteger() || new_type->isDateTime()) {
       return {builder_, expr_->cast(new_type), "", true};
     }
   }
diff --git a/omniscidb/QueryEngine/ArrowResultSetConverter.cpp b/omniscidb/QueryEngine/ArrowResultSetConverter.cpp
@@ -967,8 +967,19 @@ void append_scalar_value_and_validity(const ScalarTargetValue& value,
       }
       break;
     case hdk::ir::Type::kTime:
-      create_or_append_value<int32_t, int64_t>(value, values, max_size);
-      create_or_append_validity<int64_t>(value, type, null_bitmap, max_size);
+      switch (type->size()) {
+        case 4:
+          create_or_append_value<int32_t, int64_t>(value, values, max_size);
+          create_or_append_validity<int64_t>(value, type, null_bitmap, max_size);
+          break;
+        case 8:
+          create_or_append_value<int64_t, int64_t>(value, values, max_size);
+          create_or_append_validity<int64_t>(value, type, null_bitmap, max_size);
+          break;
+        default:
+          throw std::runtime_error(type->toString() +
+                                   " is not supported in Arrow result sets.");
+      }
       break;
     case hdk::ir::Type::kDate:
       device_type == ExecutorDeviceType::GPU
@@ -1566,8 +1577,22 @@ std::shared_ptr<arrow::DataType> get_arrow_type(const hdk::ir::Type* type,
       CHECK_EQ(type->size(), 8);
       return arrow::decimal(std::min(type->as<hdk::ir::DecimalType>()->precision(), 38),
                             type->as<hdk::ir::DecimalType>()->scale());
-    case hdk::ir::Type::kTime:
-      return time32(arrow::TimeUnit::SECOND);
+    case hdk::ir::Type::kTime: {
+      auto unit = type->as<hdk::ir::TimeType>()->unit();
+      switch (unit) {
+        case hdk::ir::TimeUnit::kSecond:
+          return time32(arrow::TimeUnit::SECOND);
+        case hdk::ir::TimeUnit::kMilli:
+          return time32(arrow::TimeUnit::MILLI);
+        case hdk::ir::TimeUnit::kMicro:
+          return time64(arrow::TimeUnit::MICRO);
+        case hdk::ir::TimeUnit::kNano:
+          return time64(arrow::TimeUnit::NANO);
+        default:
+          throw std::runtime_error("Unsupported time precision for Arrow result sets: " +
+                                   toString(unit));
+      }
+    }
     case hdk::ir::Type::kDate: {
       // TODO(wamsi) : Remove date64() once date32() support is added in cuDF. date32()
       // Currently support for date32() is missing in cuDF.Hence, if client requests for
@@ -1839,6 +1864,25 @@ void appendToColumnBuilder(ArrowResultSetConverter::ColumnBuilder& column_builde
   }
 }
 
+template <>
+void appendToColumnBuilder<arrow::Time32Builder, int64_t>(
+    ArrowResultSetConverter::ColumnBuilder& column_builder,
+    arrow::ArrayBuilder* arrow_builder,
+    const ValueArray& values,
+    const std::shared_ptr<std::vector<bool>>& is_valid) {
+  std::vector<int64_t> val_raw = boost::get<std::vector<int64_t>>(values);
+  std::vector<int32_t> vals(val_raw.begin(), val_raw.end());
+
+  auto typed_builder = dynamic_cast<arrow::Time32Builder*>(arrow_builder);
+  CHECK(typed_builder);
+  if (column_builder.field->nullable()) {
+    CHECK(is_valid.get());
+    ARROW_THROW_NOT_OK(typed_builder->AppendValues(vals, *is_valid));
+  } else {
+    ARROW_THROW_NOT_OK(typed_builder->AppendValues(vals));
+  }
+}
+
 template <>
 void appendToColumnBuilder<arrow::Decimal128Builder, int64_t>(
     ArrowResultSetConverter::ColumnBuilder& column_builder,
@@ -2005,10 +2049,25 @@ void ArrowResultSetConverter::append(
                                    " is not supported in Arrow result sets.");
       }
       break;
-    case hdk::ir::Type::kTime:
-      appendToColumnBuilder<arrow::Time32Builder, int32_t>(
-          column_builder, elem_builder, values, is_valid);
+    case hdk::ir::Type::kTime: {
+      auto unit = elem_type->as<hdk::ir::TimeType>()->unit();
+      switch (unit) {
+        case hdk::ir::TimeUnit::kSecond:
+        case hdk::ir::TimeUnit::kMilli:
+          appendToColumnBuilder<arrow::Time32Builder, int64_t>(
+              column_builder, elem_builder, values, is_valid);
+          break;
+        case hdk::ir::TimeUnit::kMicro:
+        case hdk::ir::TimeUnit::kNano:
+          appendToColumnBuilder<arrow::Time64Builder, int64_t>(
+              column_builder, elem_builder, values, is_valid);
+          break;
+        default:
+          throw std::runtime_error("Unsupported time precision for Arrow result sets: " +
+                                   toString(unit));
+      }
       break;
+    }
     case hdk::ir::Type::kTimestamp:
       appendToColumnBuilder<arrow::TimestampBuilder, int64_t>(
           column_builder, elem_builder, values, is_valid);
diff --git a/omniscidb/QueryEngine/CastIR.cpp b/omniscidb/QueryEngine/CastIR.cpp
@@ -73,58 +73,58 @@ llvm::Value* CodeGenerator::codegenCast(llvm::Value* operand_lv,
     return cgen_state_->ir_builder_.CreatePointerCast(operand_lv,
                                                       byte_array_type->getPointerTo());
   }
-  if (operand_lv->getType()->isIntegerTy()) {
-    if (operand_type->isString() || operand_type->isExtDictionary()) {
-      return codegenCastFromString(
-          operand_lv, operand_type, type, operand_is_const, is_dict_intersection, co);
-    }
-    CHECK(operand_type->isInteger() || operand_type->isDecimal() ||
-          operand_type->isDateTime() || operand_type->isBoolean());
-    if (operand_type->isBoolean()) {
-      // cast boolean to int8
-      CHECK(operand_lv->getType()->isIntegerTy(1) ||
-            operand_lv->getType()->isIntegerTy(8));
-      if (operand_lv->getType()->isIntegerTy(1)) {
-        operand_lv = cgen_state_->castToTypeIn(operand_lv, 8);
-      }
-      if (type->isBoolean()) {
-        return operand_lv;
-      }
-    }
-    if (operand_type->isInteger() && operand_lv->getType()->isIntegerTy(8) &&
-        type->isBoolean()) {
-      // cast int8 to boolean
-      return codegenCastBetweenIntTypes(operand_lv, operand_type, type);
-    }
-    if (operand_type->isTimestamp() && type->isDate()) {
-      // Maybe we should instead generate DateTruncExpr directly from RelAlgTranslator
-      // for this pattern. However, DateTruncExpr is supposed to return a timestamp,
-      // whereas this cast returns a date. The underlying type for both is still the same,
-      // but it still doesn't look like a good idea to misuse DateTruncExpr.
-      // Date will have default precision of day, but TIMESTAMP dimension would
-      // matter but while converting date through seconds
-      return codegenCastTimestampToDate(
-          operand_lv,
-          operand_type->as<hdk::ir::TimestampType>()->unit(),
-          type->nullable());
+  if (!operand_lv->getType()->isIntegerTy()) {
+    return codegenCastFromFp(operand_lv, operand_type, type);
+  }
+
+  if (operand_type->isString() || operand_type->isExtDictionary()) {
+    return codegenCastFromString(
+        operand_lv, operand_type, type, operand_is_const, is_dict_intersection, co);
+  }
+  CHECK(operand_type->isInteger() || operand_type->isDecimal() ||
+        operand_type->isDateTime() || operand_type->isBoolean());
+  if (operand_type->isBoolean()) {
+    // cast boolean to int8
+    CHECK(operand_lv->getType()->isIntegerTy(1) || operand_lv->getType()->isIntegerTy(8));
+    if (operand_lv->getType()->isIntegerTy(1)) {
+      operand_lv = cgen_state_->castToTypeIn(operand_lv, 8);
     }
-    if ((operand_type->isTimestamp() || operand_type->isDate()) && type->isTimestamp()) {
-      const auto operand_unit = (operand_type->isTimestamp())
-                                    ? operand_type->as<hdk::ir::TimestampType>()->unit()
-                                    : hdk::ir::TimeUnit::kSecond;
-      if (operand_unit != type->as<hdk::ir::TimestampType>()->unit()) {
-        return codegenCastBetweenTimestamps(
-            operand_lv, operand_type, type, type->nullable());
-      }
+    if (type->isBoolean()) {
+      return operand_lv;
     }
-    if (type->isInteger() || type->isDecimal() || type->isDateTime()) {
-      return codegenCastBetweenIntTypes(operand_lv, operand_type, type);
-    } else {
-      return codegenCastToFp(operand_lv, operand_type, type);
+  }
+  if (operand_type->isInteger() && operand_lv->getType()->isIntegerTy(8) &&
+      type->isBoolean()) {
+    // cast int8 to boolean
+    return codegenCastBetweenIntTypes(operand_lv, operand_type, type);
+  }
+  if (operand_type->isTimestamp() && type->isDate()) {
+    // Maybe we should instead generate DateTruncExpr directly from RelAlgTranslator
+    // for this pattern. However, DateTruncExpr is supposed to return a timestamp,
+    // whereas this cast returns a date. The underlying type for both is still the same,
+    // but it still doesn't look like a good idea to misuse DateTruncExpr.
+    // Date will have default precision of day, but TIMESTAMP dimension would
+    // matter but while converting date through seconds
+    return codegenCastTimestampToDate(
+        operand_lv, operand_type->as<hdk::ir::TimestampType>()->unit(), type->nullable());
+  }
+  if (operand_type->isTimestamp() && type->isTime()) {
+    return codegenCastTimestampToTime(operand_lv, operand_type, type);
+  }
+  if ((operand_type->isTimestamp() || operand_type->isDate()) && type->isTimestamp()) {
+    const auto operand_unit = (operand_type->isTimestamp())
+                                  ? operand_type->as<hdk::ir::TimestampType>()->unit()
+                                  : hdk::ir::TimeUnit::kSecond;
+    if (operand_unit != type->as<hdk::ir::TimestampType>()->unit()) {
+      return codegenCastBetweenTimestamps(operand_lv, operand_type, type);
     }
+  }
+  if (type->isInteger() || type->isDecimal() || type->isDateTime()) {
+    return codegenCastBetweenIntTypes(operand_lv, operand_type, type);
   } else {
-    return codegenCastFromFp(operand_lv, operand_type, type);
+    return codegenCastToFp(operand_lv, operand_type, type);
   }
+
   CHECK(false);
   return nullptr;
 }
@@ -159,11 +159,53 @@ llvm::Value* CodeGenerator::codegenCastTimestampToDate(llvm::Value* ts_lv,
   return ret;
 }
 
+llvm::Value* CodeGenerator::codegenCastTimestampToTime(llvm::Value* ts_lv,
+                                                       const hdk::ir::Type* operand_type,
+                                                       const hdk::ir::Type* target_type) {
+  AUTOMATIC_IR_METADATA(cgen_state_);
+  const auto operand_unit = operand_type->as<hdk::ir::TimestampType>()->unit();
+  const auto target_unit = target_type->as<hdk::ir::TimeType>()->unit();
+
+  CHECK(ts_lv->getType()->isIntegerTy(64));
+  if (operand_unit < target_unit) {
+    const auto target_sec_scale = hdk::ir::unitsPerSecond(target_unit);
+    const auto scale =
+        hdk::ir::unitsPerSecond(target_unit) / hdk::ir::unitsPerSecond(operand_unit);
+    if (!operand_type->nullable()) {
+      return cgen_state_->emitCall(
+          "TimestampToTimeUpscale",
+          {{ts_lv, cgen_state_->llInt(scale), cgen_state_->llInt(target_sec_scale)}});
+    }
+    // Timestamp, Time32. Time64 have int64 inline int value, but logically corect to pass
+    // target and source null values.
+    return cgen_state_->emitCall("TimestampToTimeUpscaleNullable",
+                                 {{ts_lv,
+                                   cgen_state_->llInt(scale),
+                                   cgen_state_->llInt(target_sec_scale),
+                                   cgen_state_->inlineIntNull(operand_type),
+                                   cgen_state_->inlineIntNull(target_type)}});
+  }
+  // On downscale we at first cut of date part, than cut of unused unit precision
+  const auto operand_sec_scale = hdk::ir::unitsPerSecond(operand_unit);
+  const auto scale =
+      hdk::ir::unitsPerSecond(operand_unit) / hdk::ir::unitsPerSecond(target_unit);
+  if (!operand_type->nullable()) {
+    return cgen_state_->emitCall(
+        "TimestampToTimeDownscale",
+        {{ts_lv, cgen_state_->llInt(scale), cgen_state_->llInt(operand_sec_scale)}});
+  }
+  return cgen_state_->emitCall("TimestampToTimeDownscaleNullable",
+                               {{ts_lv,
+                                 cgen_state_->llInt(scale),
+                                 cgen_state_->llInt(operand_sec_scale),
+                                 cgen_state_->inlineIntNull(operand_type),
+                                 cgen_state_->inlineIntNull(target_type)}});
+}
+
 llvm::Value* CodeGenerator::codegenCastBetweenTimestamps(
     llvm::Value* ts_lv,
     const hdk::ir::Type* operand_type,
-    const hdk::ir::Type* target_type,
-    const bool nullable) {
+    const hdk::ir::Type* target_type) {
   AUTOMATIC_IR_METADATA(cgen_state_);
   const auto operand_unit = operand_type->isTimestamp()
                                 ? operand_type->as<hdk::ir::TimestampType>()->unit()
@@ -179,7 +221,7 @@ llvm::Value* CodeGenerator::codegenCastBetweenTimestamps(
     const auto scale =
         hdk::ir::unitsPerSecond(target_unit) / hdk::ir::unitsPerSecond(operand_unit);
     codegenCastBetweenIntTypesOverflowChecks(ts_lv, operand_type, target_type, scale);
-    return nullable
+    return target_type->nullable()
                ? cgen_state_->emitCall("mul_int64_t_nullable_lhs",
                                        {ts_lv,
                                         cgen_state_->llInt(static_cast<int64_t>(scale)),
@@ -189,7 +231,7 @@ llvm::Value* CodeGenerator::codegenCastBetweenTimestamps(
   }
   const auto scale =
       hdk::ir::unitsPerSecond(operand_unit) / hdk::ir::unitsPerSecond(target_unit);
-  return nullable
+  return target_type->nullable()
              ? cgen_state_->emitCall("floor_div_nullable_lhs",
                                      {ts_lv,
                                       cgen_state_->llInt(static_cast<int64_t>(scale)),
diff --git a/omniscidb/QueryEngine/CodeGenerator.h b/omniscidb/QueryEngine/CodeGenerator.h
@@ -265,10 +265,13 @@ class CodeGenerator {
                                           const hdk::ir::TimeUnit unit,
                                           const bool nullable);
 
+  llvm::Value* codegenCastTimestampToTime(llvm::Value* ts_lv,
+                                          const hdk::ir::Type* operand_type,
+                                          const hdk::ir::Type* target_type);
+
   llvm::Value* codegenCastBetweenTimestamps(llvm::Value* ts_lv,
                                             const hdk::ir::Type* operand_type,
-                                            const hdk::ir::Type* target_type,
-                                            const bool nullable);
+                                            const hdk::ir::Type* target_type);
 
   llvm::Value* codegenCastFromString(llvm::Value* operand_lv,
                                      const hdk::ir::Type* operand_type,
diff --git a/omniscidb/QueryEngine/DateTruncate.cpp b/omniscidb/QueryEngine/DateTruncate.cpp
@@ -330,6 +330,66 @@ DateTruncateHighPrecisionToDateNullable(const int64_t timeval,
   return truncate_high_precision_timestamp_to_date(timeval, scale);
 }
 
+namespace {
+// scale is 10^{3,6,9}
+inline DEVICE int64_t
+timestamp_to_time_truncation_upscale(const int64_t timeval,
+                                     const int64_t conversion_scale,
+                                     const int64_t target_sec_scale) {
+  return unsigned_mod(timeval * conversion_scale, target_sec_scale * kSecsPerDay);
+}
+
+// scale is 10^{3,6,9}
+inline DEVICE int64_t
+timestamp_to_time_truncation_downscale(const int64_t timeval,
+                                       const int64_t conversion_scale,
+                                       const int64_t target_sec_scale) {
+  return unsigned_mod(timeval, target_sec_scale * kSecsPerDay) / conversion_scale;
+}
+}  // namespace
+
+extern "C" DEVICE RUNTIME_EXPORT ALWAYS_INLINE int64_t
+TimestampToTimeUpscale(const int64_t timeval,
+                       const int64_t conversion_scale,
+                       const int64_t target_sec_scale) {
+  return timestamp_to_time_truncation_upscale(
+      timeval, conversion_scale, target_sec_scale);
+}
+
+extern "C" DEVICE RUNTIME_EXPORT ALWAYS_INLINE int64_t
+TimestampToTimeDownscale(const int64_t timeval,
+                         const int64_t conversion_scale,
+                         const int64_t target_sec_scale) {
+  return timestamp_to_time_truncation_downscale(
+      timeval, conversion_scale, target_sec_scale);
+}
+
+extern "C" RUNTIME_EXPORT ALWAYS_INLINE DEVICE int64_t
+TimestampToTimeUpscaleNullable(const int64_t timeval,
+                               const int64_t conversion_scale,
+                               const int64_t target_sec_scale,
+                               const int64_t source_null_val,
+                               const int64_t target_null_val) {
+  if (timeval == source_null_val) {
+    return target_null_val;
+  }
+  return timestamp_to_time_truncation_upscale(
+      timeval, conversion_scale, target_sec_scale);
+}
+
+extern "C" RUNTIME_EXPORT ALWAYS_INLINE DEVICE int64_t
+TimestampToTimeDownscaleNullable(const int64_t timeval,
+                                 const int64_t conversion_scale,
+                                 const int64_t target_sec_scale,
+                                 const int64_t source_null_val,
+                                 const int64_t target_null_val) {
+  if (timeval == source_null_val) {
+    return target_null_val;
+  }
+  return timestamp_to_time_truncation_downscale(
+      timeval, conversion_scale, target_sec_scale);
+}
+
 namespace {
 
 struct EraTime {
diff --git a/omniscidb/ResultSet/ResultSet.cpp b/omniscidb/ResultSet/ResultSet.cpp
diff --git a/omniscidb/Tests/ArrowTestHelpers.h b/omniscidb/Tests/ArrowTestHelpers.h
diff --git a/omniscidb/Tests/QueryBuilderTest.cpp b/omniscidb/Tests/QueryBuilderTest.cpp

Original file line number	Diff line number	Diff line change
`@@ -915,11 +915,11 @@ BuilderExpr BuilderExpr::cast(const Type* new_type) const {`
`915`	`915`	`return {builder_, expr_->cast(new_type), "", true};`
`916`	`916`	`}`
`917`	`917`	`} else if (expr_->type()->isTime()) {`
`918`		`- if (new_type->isTime()) {`
	`918`	`+ if (new_type->isInteger() \|\| new_type->isTime()) {`
`919`	`919`	`return {builder_, expr_->cast(new_type), "", true};`
`920`	`920`	`}`
`921`	`921`	`} else if (expr_->type()->isTimestamp()) {`
`922`		`- if (new_type->isNumber() \|\| new_type->isDate() \|\| new_type->isTimestamp()) {`
	`922`	`+ if (new_type->isInteger() \|\| new_type->isDateTime()) {`
`923`	`923`	`return {builder_, expr_->cast(new_type), "", true};`
`924`	`924`	`}`
`925`	`925`	`}`