The Gandiva Expression Compiler

Gandiva is a runtime expression compiler that uses LLVM to generate efficient native code for projections and filters on Arrow record batches. Gandiva only handles projections and filters. For other transformation, see Compute Functions.

Gandiva was designed to take advantage of the Arrow memory format and modern hardware. Compiling expressions using LLVM allows the execution to be optimized to the local runtime environment and hardware, including available SIMD instructions. To minimize optimization overhead, all Gandiva functions are pre-compiled into LLVM IR (intermediate representation).

Building Expressions

Gandiva provides a general expression representation where expressions are represented by a tree of nodes. The expression trees are built using gandiva::TreeExprBuilder. The leaves of the expression tree are typically field references, created by gandiva::TreeExprBuilder::MakeField(), and literal values, created by gandiva::TreeExprBuilder::MakeLiteral(). Nodes can be combined into more complex expression trees using:

• gandiva::TreeExprBuilder::MakeFunction() to create a function node. See available functions below.

• gandiva::TreeExprBuilder::MakeIf() to create if-else logic.

• gandiva::TreeExprBuilder::MakeAnd() and gandiva::TreeExprBuilder::MakeOr() to create boolean expressions. (For “not”, use the not(bool) function in MakeFunction.)

• gandiva::TreeExprBuilder::MakeInExpressionInt32() and the other “in expression” functions to create set membership tests.

Once an expression tree is built, they are wrapped in either gandiva::Expression or gandiva::Condition, depending on how they will be used. Expression is used in projections while Condition is used filters.

As an example, here is how to create an Expression representing x + 3 and a Condition representing x < 3:

auto field_x_raw = arrow::field("x", arrow::int32());
auto field_x = TreeExprBuilder::MakeField(field_x_raw);
auto literal_3 = TreeExprBuilder::MakeLiteral(3);
auto field_result = arrow::field("result", arrow::int32());

auto add_node = TreeExprBuilder::MakeFunction("add", {field_x, literal_3}, arrow::int32());
auto expression = TreeExprBuilder::MakeExpression(add_node, field_result);

auto less_than_node = TreeExprBuilder::MakeFunction("less_than", {field_x, literal_3},
                                                    boolean());
auto condition = TreeExprBuilder::MakeCondition(less_than_node);

For simpler expressions, there are also convenience functions that allow you to use functions directly in MakeExpression and MakeCondition:

auto expression = TreeExprBuilder::MakeExpression("add", {field_x, literal_3}, field_result);

auto condition = TreeExprBuilder::MakeCondition("less_than", {field_x, literal_3});

Projectors and Filters

Gandiva’s two execution kernels are gandiva::Projector and gandiva::Filter. Projector consumes a record batch and projects into a new record batch. Filter consumes a record batch and produces a gandiva::SelectionVector containing the indices that matched the condition.

For both Projector and Filter, optimization of the expression IR happens when creating instances. They are compiled against a static schema, so the schema of the record batches must be known at this point.

Continuing with the expression and condition created in the previous section, here is an example of creating a Projector and a Filter:

auto schema = arrow::schema({field_x});
std::shared_ptr<Projector> projector;
auto status = Projector::Make(schema, {expression}, &projector);
ARROW_CHECK_OK(status);

std::shared_ptr<Filter> filter;
status = Filter::Make(schema, condition, &filter);
ARROW_CHECK_OK(status);

Once a Projector or Filter is created, it can be evaluated on Arrow record batches. These execution kernels are single-threaded on their own, but are designed to be reused to process record batches in parallel.

Execution is performed with gandiva::Projector::Evaluate() and gandiva::Filter::Evaluate(). Filters produce gandiva::SelectionVector, a vector of row indices that matched the filter condition. When filtering and projecting record batches, you can pass the selection vector into the projector so that the projection is only evaluated on matching rows.

Here is an example of evaluating the Filter and Projector created above:

auto pool = arrow::default_memory_pool();
int num_records = 4;
auto array = MakeArrowArrayInt32({1, 2, 3, 4}, {true, true, true, true});
auto in_batch = arrow::RecordBatch::Make(schema, num_records, {array});

// Just project
arrow::ArrayVector outputs;
status = projector->Evaluate(*in_batch, pool, &outputs);
ARROW_CHECK_OK(status);

// Evaluate filter
gandiva::SelectionVector result_indices;
status = filter->Evaluate(*in_batch, &result_indices);
ARROW_CHECK_OK(status);

// Project with filter
arrow::ArrayVector outputs_filtered;
status = projector->Evaluate(*in_batch, selection_vector.get(),
                             pool, &outputs_filtered);

Available Gandiva Functions

Comparisons

not	(bool) -> bool
equal	(int8, int8) -> bool
	(int16, int16) -> bool
	(int32, int32) -> bool
	(int64, int64) -> bool
	(uint8, uint8) -> bool
	(uint16, uint16) -> bool
	(uint32, uint32) -> bool
	(uint64, uint64) -> bool
	(float, float) -> bool
	(double, double) -> bool
	(decimal128(38, 0), decimal128(38, 0)) -> bool
	(date64[ms], date64[ms]) -> bool
	(timestamp[ms], timestamp[ms]) -> bool
	(time32[ms], time32[ms]) -> bool
	(date32[day], date32[day]) -> bool
	(bool, bool) -> bool
	(string, string) -> bool
	(binary, binary) -> bool
eq	(int8, int8) -> bool
	(int16, int16) -> bool
	(int32, int32) -> bool
	(int64, int64) -> bool
	(uint8, uint8) -> bool
	(uint16, uint16) -> bool
	(uint32, uint32) -> bool
	(uint64, uint64) -> bool
	(float, float) -> bool
	(double, double) -> bool
	(decimal128(38, 0), decimal128(38, 0)) -> bool
	(date64[ms], date64[ms]) -> bool
	(timestamp[ms], timestamp[ms]) -> bool
	(time32[ms], time32[ms]) -> bool
	(date32[day], date32[day]) -> bool
	(bool, bool) -> bool
same	(int8, int8) -> bool
	(int16, int16) -> bool
	(int32, int32) -> bool
	(int64, int64) -> bool
	(uint8, uint8) -> bool
	(uint16, uint16) -> bool
	(uint32, uint32) -> bool
	(uint64, uint64) -> bool
	(float, float) -> bool
	(double, double) -> bool
	(decimal128(38, 0), decimal128(38, 0)) -> bool
	(date64[ms], date64[ms]) -> bool
	(timestamp[ms], timestamp[ms]) -> bool
	(time32[ms], time32[ms]) -> bool
	(date32[day], date32[day]) -> bool
	(bool, bool) -> bool
not_equal	(int8, int8) -> bool
	(int16, int16) -> bool
	(int32, int32) -> bool
	(int64, int64) -> bool
	(uint8, uint8) -> bool
	(uint16, uint16) -> bool
	(uint32, uint32) -> bool
	(uint64, uint64) -> bool
	(float, float) -> bool
	(double, double) -> bool
	(decimal128(38, 0), decimal128(38, 0)) -> bool
	(date64[ms], date64[ms]) -> bool
	(timestamp[ms], timestamp[ms]) -> bool
	(time32[ms], time32[ms]) -> bool
	(date32[day], date32[day]) -> bool
	(bool, bool) -> bool
	(string, string) -> bool
	(binary, binary) -> bool
less_than	(int8, int8) -> bool
	(int16, int16) -> bool
	(int32, int32) -> bool
	(int64, int64) -> bool
	(uint8, uint8) -> bool
	(uint16, uint16) -> bool
	(uint32, uint32) -> bool
	(uint64, uint64) -> bool
	(float, float) -> bool
	(double, double) -> bool
	(decimal128(38, 0), decimal128(38, 0)) -> bool
	(date64[ms], date64[ms]) -> bool
	(timestamp[ms], timestamp[ms]) -> bool
	(time32[ms], time32[ms]) -> bool
	(date32[day], date32[day]) -> bool
	(string, string) -> bool
	(binary, binary) -> bool
less_than_or_equal_to	(int8, int8) -> bool
	(int16, int16) -> bool
	(int32, int32) -> bool
	(int64, int64) -> bool
	(uint8, uint8) -> bool
	(uint16, uint16) -> bool
	(uint32, uint32) -> bool
	(uint64, uint64) -> bool
	(float, float) -> bool
	(double, double) -> bool
	(decimal128(38, 0), decimal128(38, 0)) -> bool
	(date64[ms], date64[ms]) -> bool
	(timestamp[ms], timestamp[ms]) -> bool
	(time32[ms], time32[ms]) -> bool
	(date32[day], date32[day]) -> bool
	(string, string) -> bool
	(binary, binary) -> bool
greater_than	(int8, int8) -> bool
	(int16, int16) -> bool
	(int32, int32) -> bool
	(int64, int64) -> bool
	(uint8, uint8) -> bool
	(uint16, uint16) -> bool
	(uint32, uint32) -> bool
	(uint64, uint64) -> bool
	(float, float) -> bool
	(double, double) -> bool
	(decimal128(38, 0), decimal128(38, 0)) -> bool
	(date64[ms], date64[ms]) -> bool
	(timestamp[ms], timestamp[ms]) -> bool
	(time32[ms], time32[ms]) -> bool
	(date32[day], date32[day]) -> bool
	(string, string) -> bool
	(binary, binary) -> bool
greater_than_or_equal_to	(int8, int8) -> bool
	(int16, int16) -> bool
	(int32, int32) -> bool
	(int64, int64) -> bool
	(uint8, uint8) -> bool
	(uint16, uint16) -> bool
	(uint32, uint32) -> bool
	(uint64, uint64) -> bool
	(float, float) -> bool
	(double, double) -> bool
	(decimal128(38, 0), decimal128(38, 0)) -> bool
	(date64[ms], date64[ms]) -> bool
	(timestamp[ms], timestamp[ms]) -> bool
	(time32[ms], time32[ms]) -> bool
	(date32[day], date32[day]) -> bool
	(string, string) -> bool
	(binary, binary) -> bool
isnull	(int8) -> bool
	(int16) -> bool
	(int32) -> bool
	(int64) -> bool
	(uint8) -> bool
	(uint16) -> bool
	(uint32) -> bool
	(uint64) -> bool
	(float) -> bool
	(double) -> bool
	(decimal128(38, 0)) -> bool
	(date64[ms]) -> bool
	(timestamp[ms]) -> bool
	(time32[ms]) -> bool
	(date32[day]) -> bool
	(bool) -> bool
	(string) -> bool
	(binary) -> bool
isnotnull	(int8) -> bool
	(int16) -> bool
	(int32) -> bool
	(int64) -> bool
	(uint8) -> bool
	(uint16) -> bool
	(uint32) -> bool
	(uint64) -> bool
	(float) -> bool
	(double) -> bool
	(decimal128(38, 0)) -> bool
	(date64[ms]) -> bool
	(timestamp[ms]) -> bool
	(time32[ms]) -> bool
	(date32[day]) -> bool
	(bool) -> bool
	(string) -> bool
	(binary) -> bool
isnumeric	(int8) -> bool
	(int16) -> bool
	(int32) -> bool
	(int64) -> bool
	(uint8) -> bool
	(uint16) -> bool
	(uint32) -> bool
	(uint64) -> bool
	(float) -> bool
	(double) -> bool
	(decimal128(38, 0)) -> bool
is_distinct_from	(int8, int8) -> bool
	(int16, int16) -> bool
	(int32, int32) -> bool
	(int64, int64) -> bool
	(uint8, uint8) -> bool
	(uint16, uint16) -> bool
	(uint32, uint32) -> bool
	(uint64, uint64) -> bool
	(float, float) -> bool
	(double, double) -> bool
	(decimal128(38, 0), decimal128(38, 0)) -> bool
	(date64[ms], date64[ms]) -> bool
	(timestamp[ms], timestamp[ms]) -> bool
	(time32[ms], time32[ms]) -> bool
	(date32[day], date32[day]) -> bool
	(bool, bool) -> bool
is_not_distinct_from	(int8, int8) -> bool
	(int16, int16) -> bool
	(int32, int32) -> bool
	(int64, int64) -> bool
	(uint8, uint8) -> bool
	(uint16, uint16) -> bool
	(uint32, uint32) -> bool
	(uint64, uint64) -> bool
	(float, float) -> bool
	(double, double) -> bool
	(decimal128(38, 0), decimal128(38, 0)) -> bool
	(date64[ms], date64[ms]) -> bool
	(timestamp[ms], timestamp[ms]) -> bool
	(time32[ms], time32[ms]) -> bool
	(date32[day], date32[day]) -> bool
	(bool, bool) -> bool

Cast

castBIGINT	(int32) -> int64
	(decimal128(38, 0)) -> int64
	(day_time_interval) -> int64
	(string) -> int64
castINT	(int64) -> int32
	(string) -> int32
castFLOAT4	(int32) -> float
	(int64) -> float
	(double) -> float
	(string) -> float
castFLOAT8	(int32) -> double
	(int64) -> double
	(float) -> double
	(decimal128(38, 0)) -> double
	(string) -> double
castDECIMAL	(int32) -> decimal128(38, 0)
	(int64) -> decimal128(38, 0)
	(float) -> decimal128(38, 0)
	(double) -> decimal128(38, 0)
	(decimal128(38, 0)) -> decimal128(38, 0)
	(string) -> decimal128(38, 0)
castDECIMALNullOnOverflow	(decimal128(38, 0)) -> decimal128(38, 0)
castDATE	(int64) -> date64[ms]
	(int32) -> date32[day]
	(date32[day]) -> date64[ms]
	(string) -> date64[ms]
	(timestamp[ms]) -> date64[ms]
castTIMESTAMP	(string) -> timestamp[ms]
	(date64[ms]) -> timestamp[ms]
	(int64) -> timestamp[ms]
castVARCHAR	(timestamp[ms], int64) -> string
	(bool, int64) -> string
	(string, int64) -> string
	(binary, int64) -> string
	(int32, int64) -> string
	(int64, int64) -> string
	(float, int64) -> string
	(double, int64) -> string
	(decimal128(38, 0), int64) -> string
castTIME	(timestamp[ms]) -> time32[ms]
castBIT	(string) -> bool
castBOOLEAN	(string) -> bool
castVARBINARY	(binary, int64) -> binary
	(string, int64) -> binary
	(int32, int64) -> binary
	(int64, int64) -> binary
	(float, int64) -> binary
	(double, int64) -> binary

Arithmetic and Math

add	(int8, int8) -> int8
	(int16, int16) -> int16
	(int32, int32) -> int32
	(int64, int64) -> int64
	(uint8, uint8) -> uint8
	(uint16, uint16) -> uint16
	(uint32, uint32) -> uint32
	(uint64, uint64) -> uint64
	(float, float) -> float
	(double, double) -> double
	(decimal128(38, 0), decimal128(38, 0)) -> decimal128(38, 0)
	(date64[ms], int32) -> date64[ms]
	(timestamp[ms], int32) -> timestamp[ms]
	(date64[ms], int64) -> date64[ms]
	(timestamp[ms], int64) -> timestamp[ms]
	(int32, date64[ms]) -> date64[ms]
	(int32, timestamp[ms]) -> timestamp[ms]
	(int64, date64[ms]) -> date64[ms]
	(int64, timestamp[ms]) -> timestamp[ms]
	(date64[ms], int64) -> timestamp[ms]
subtract	(int8, int8) -> int8
	(int16, int16) -> int16
	(int32, int32) -> int32
	(int64, int64) -> int64
	(uint8, uint8) -> uint8
	(uint16, uint16) -> uint16
	(uint32, uint32) -> uint32
	(uint64, uint64) -> uint64
	(float, float) -> float
	(double, double) -> double
	(decimal128(38, 0), decimal128(38, 0)) -> decimal128(38, 0)
	(date64[ms], int32) -> date64[ms]
	(timestamp[ms], int32) -> timestamp[ms]
	(date64[ms], int64) -> date64[ms]
	(timestamp[ms], int64) -> timestamp[ms]
multiply	(int8, int8) -> int8
	(int16, int16) -> int16
	(int32, int32) -> int32
	(int64, int64) -> int64
	(uint8, uint8) -> uint8
	(uint16, uint16) -> uint16
	(uint32, uint32) -> uint32
	(uint64, uint64) -> uint64
	(float, float) -> float
	(double, double) -> double
	(decimal128(38, 0), decimal128(38, 0)) -> decimal128(38, 0)
divide	(int8, int8) -> int8
	(int16, int16) -> int16
	(int32, int32) -> int32
	(int64, int64) -> int64
	(uint8, uint8) -> uint8
	(uint16, uint16) -> uint16
	(uint32, uint32) -> uint32
	(uint64, uint64) -> uint64
	(float, float) -> float
	(double, double) -> double
	(decimal128(38, 0), decimal128(38, 0)) -> decimal128(38, 0)
mod	(int64, int32) -> int32
	(int64, int64) -> int64
	(decimal128(38, 0), decimal128(38, 0)) -> decimal128(38, 0)
	(double, double) -> double
modulo	(int64, int32) -> int32
	(int64, int64) -> int64
	(decimal128(38, 0), decimal128(38, 0)) -> decimal128(38, 0)
	(double, double) -> double
div	(int32, int32) -> int32
	(int64, int64) -> int64
	(float, float) -> float
	(double, double) -> double
bitwise_and	(int32, int32) -> int32
	(int64, int64) -> int64
bitwise_or	(int32, int32) -> int32
	(int64, int64) -> int64
bitwise_xor	(int32, int32) -> int32
	(int64, int64) -> int64
bitwise_not	(int32) -> int32
	(int64) -> int64
round	(float) -> float
	(double) -> double
	(float, int32) -> float
	(double, int32) -> double
	(int32) -> int32
	(int64) -> int64
	(int32, int32) -> int32
	(int64, int32) -> int64
	(decimal128(38, 0)) -> decimal128(38, 0)
	(decimal128(38, 0), int32) -> decimal128(38, 0)
cbrt	(int32) -> double
	(int64) -> double
	(uint32) -> double
	(uint64) -> double
	(float) -> double
	(double) -> double
exp	(int32) -> double
	(int64) -> double
	(uint32) -> double
	(uint64) -> double
	(float) -> double
	(double) -> double
log	(int32) -> double
	(int64) -> double
	(uint32) -> double
	(uint64) -> double
	(float) -> double
	(double) -> double
	(int32, int32) -> double
	(int64, int64) -> double
	(uint32, uint32) -> double
	(uint64, uint64) -> double
	(float, float) -> double
	(double, double) -> double
log10	(int32) -> double
	(int64) -> double
	(uint32) -> double
	(uint64) -> double
	(float) -> double
	(double) -> double
power	(double, double) -> double
pow	(double, double) -> double
sin	(int32) -> double
	(int64) -> double
	(uint32) -> double
	(uint64) -> double
	(float) -> double
	(double) -> double
cos	(int32) -> double
	(int64) -> double
	(uint32) -> double
	(uint64) -> double
	(float) -> double
	(double) -> double
asin	(int32) -> double
	(int64) -> double
	(uint32) -> double
	(uint64) -> double
	(float) -> double
	(double) -> double
acos	(int32) -> double
	(int64) -> double
	(uint32) -> double
	(uint64) -> double
	(float) -> double
	(double) -> double
tan	(int32) -> double
	(int64) -> double
	(uint32) -> double
	(uint64) -> double
	(float) -> double
	(double) -> double
atan	(int32) -> double
	(int64) -> double
	(uint32) -> double
	(uint64) -> double
	(float) -> double
	(double) -> double
sinh	(int32) -> double
	(int64) -> double
	(uint32) -> double
	(uint64) -> double
	(float) -> double
	(double) -> double
cosh	(int32) -> double
	(int64) -> double
	(uint32) -> double
	(uint64) -> double
	(float) -> double
	(double) -> double
tanh	(int32) -> double
	(int64) -> double
	(uint32) -> double
	(uint64) -> double
	(float) -> double
	(double) -> double
cot	(int32) -> double
	(int64) -> double
	(uint32) -> double
	(uint64) -> double
	(float) -> double
	(double) -> double
radians	(int32) -> double
	(int64) -> double
	(uint32) -> double
	(uint64) -> double
	(float) -> double
	(double) -> double
degrees	(int32) -> double
	(int64) -> double
	(uint32) -> double
	(uint64) -> double
	(float) -> double
	(double) -> double
atan2	(int32, int32) -> double
	(int64, int64) -> double
	(uint32, uint32) -> double
	(uint64, uint64) -> double
	(float, float) -> double
	(double, double) -> double
abs	(decimal128(38, 0)) -> decimal128(38, 0)
ceil	(decimal128(38, 0)) -> decimal128(38, 0)
floor	(decimal128(38, 0)) -> decimal128(38, 0)
truncate	(decimal128(38, 0)) -> decimal128(38, 0)
	(decimal128(38, 0), int32) -> decimal128(38, 0)
	(int64, int32) -> int64
trunc	(decimal128(38, 0)) -> decimal128(38, 0)
	(decimal128(38, 0), int32) -> decimal128(38, 0)
	(int64, int32) -> int64
random	() -> double
	(int32) -> double
rand	() -> double
	(int32) -> double

Other

bin	(int32) -> string
	(int64) -> string
to_time	(int8) -> time32[ms]
	(int16) -> time32[ms]
	(int32) -> time32[ms]
	(int64) -> time32[ms]
	(uint8) -> time32[ms]
	(uint16) -> time32[ms]
	(uint32) -> time32[ms]
	(uint64) -> time32[ms]
	(float) -> time32[ms]
	(double) -> time32[ms]
to_timestamp	(int8) -> timestamp[ms]
	(int16) -> timestamp[ms]
	(int32) -> timestamp[ms]
	(int64) -> timestamp[ms]
	(uint8) -> timestamp[ms]
	(uint16) -> timestamp[ms]
	(uint32) -> timestamp[ms]
	(uint64) -> timestamp[ms]
	(float) -> timestamp[ms]
	(double) -> timestamp[ms]
sha	(int8) -> string
	(int16) -> string
	(int32) -> string
	(int64) -> string
	(uint8) -> string
	(uint16) -> string
	(uint32) -> string
	(uint64) -> string
	(float) -> string
	(double) -> string
	(decimal128(38, 0)) -> string
	(bool) -> string
sha1	(int8) -> string
	(int16) -> string
	(int32) -> string
	(int64) -> string
	(uint8) -> string
	(uint16) -> string
	(uint32) -> string
	(uint64) -> string
	(float) -> string
	(double) -> string
	(decimal128(38, 0)) -> string
	(bool) -> string
sha256	(int8) -> string
	(int16) -> string
	(int32) -> string
	(int64) -> string
	(uint8) -> string
	(uint16) -> string
	(uint32) -> string
	(uint64) -> string
	(float) -> string
	(double) -> string
	(decimal128(38, 0)) -> string
	(bool) -> string
space	(int32) -> string
	(int64) -> string
convert_toDOUBLE	(double) -> binary
convert_toDOUBLE_be	(double) -> binary
convert_toFLOAT	(float) -> binary
convert_toFLOAT_be	(float) -> binary
convert_toINT	(int32) -> binary
convert_toINT_be	(int32) -> binary
convert_toBIGINT	(int64) -> binary
convert_toBIGINT_be	(int64) -> binary
convert_toBOOLEAN_BYTE	(bool) -> binary

Dates and Timestamps

extractMillennium	(date64[ms]) -> int64
	(timestamp[ms]) -> int64
extractCentury	(date64[ms]) -> int64
	(timestamp[ms]) -> int64
extractDecade	(date64[ms]) -> int64
	(timestamp[ms]) -> int64
extractYear	(date64[ms]) -> int64
	(timestamp[ms]) -> int64
year	(date64[ms]) -> int64
	(timestamp[ms]) -> int64
	(date64[ms]) -> date64[ms]
	(timestamp[ms]) -> timestamp[ms]
extractQuarter	(date64[ms]) -> int64
	(timestamp[ms]) -> int64
extractMonth	(date64[ms]) -> int64
	(timestamp[ms]) -> int64
month	(date64[ms]) -> int64
	(timestamp[ms]) -> int64
	(date64[ms]) -> date64[ms]
	(timestamp[ms]) -> timestamp[ms]
extractWeek	(date64[ms]) -> int64
	(timestamp[ms]) -> int64
weekofyear	(date64[ms]) -> int64
	(timestamp[ms]) -> int64
	(date64[ms]) -> date64[ms]
	(timestamp[ms]) -> timestamp[ms]
yearweek	(date64[ms]) -> int64
	(timestamp[ms]) -> int64
	(date64[ms]) -> date64[ms]
	(timestamp[ms]) -> timestamp[ms]
extractDay	(date64[ms]) -> int64
	(timestamp[ms]) -> int64
	(day_time_interval) -> int64
day	(date64[ms]) -> int64
	(timestamp[ms]) -> int64
	(date64[ms]) -> date64[ms]
	(timestamp[ms]) -> timestamp[ms]
dayofmonth	(date64[ms]) -> int64
	(timestamp[ms]) -> int64
	(date64[ms]) -> date64[ms]
	(timestamp[ms]) -> timestamp[ms]
extractHour	(date64[ms]) -> int64
	(timestamp[ms]) -> int64
	(time32[ms]) -> int64
hour	(date64[ms]) -> int64
	(timestamp[ms]) -> int64
	(date64[ms]) -> date64[ms]
	(timestamp[ms]) -> timestamp[ms]
	(time32[ms]) -> int64
extractMinute	(date64[ms]) -> int64
	(timestamp[ms]) -> int64
	(time32[ms]) -> int64
minute	(date64[ms]) -> int64
	(timestamp[ms]) -> int64
	(date64[ms]) -> date64[ms]
	(timestamp[ms]) -> timestamp[ms]
	(time32[ms]) -> int64
extractSecond	(date64[ms]) -> int64
	(timestamp[ms]) -> int64
	(time32[ms]) -> int64
second	(date64[ms]) -> int64
	(timestamp[ms]) -> int64
	(date64[ms]) -> date64[ms]
	(timestamp[ms]) -> timestamp[ms]
	(time32[ms]) -> int64
date_trunc_Millennium	(date64[ms]) -> date64[ms]
	(timestamp[ms]) -> timestamp[ms]
date_trunc_Century	(date64[ms]) -> date64[ms]
	(timestamp[ms]) -> timestamp[ms]
date_trunc_Decade	(date64[ms]) -> date64[ms]
	(timestamp[ms]) -> timestamp[ms]
date_trunc_Year	(date64[ms]) -> date64[ms]
	(timestamp[ms]) -> timestamp[ms]
date_trunc_Quarter	(date64[ms]) -> date64[ms]
	(timestamp[ms]) -> timestamp[ms]
date_trunc_Month	(date64[ms]) -> date64[ms]
	(timestamp[ms]) -> timestamp[ms]
date_trunc_Week	(date64[ms]) -> date64[ms]
	(timestamp[ms]) -> timestamp[ms]
date_trunc_Day	(date64[ms]) -> date64[ms]
	(timestamp[ms]) -> timestamp[ms]
date_trunc_Hour	(date64[ms]) -> date64[ms]
	(timestamp[ms]) -> timestamp[ms]
date_trunc_Minute	(date64[ms]) -> date64[ms]
	(timestamp[ms]) -> timestamp[ms]
date_trunc_Second	(date64[ms]) -> date64[ms]
	(timestamp[ms]) -> timestamp[ms]
extractDoy	(date64[ms]) -> int64
	(timestamp[ms]) -> int64
extractDow	(date64[ms]) -> int64
	(timestamp[ms]) -> int64
extractEpoch	(date64[ms]) -> int64
	(timestamp[ms]) -> int64
to_date	(timestamp[ms]) -> date64[ms]
last_day	(date64[ms]) -> date64[ms]
	(timestamp[ms]) -> date64[ms]
sha	(date64[ms]) -> string
	(timestamp[ms]) -> string
	(time32[ms]) -> string
	(date32[day]) -> string
sha1	(date64[ms]) -> string
	(timestamp[ms]) -> string
	(time32[ms]) -> string
	(date32[day]) -> string
sha256	(date64[ms]) -> string
	(timestamp[ms]) -> string
	(time32[ms]) -> string
	(date32[day]) -> string
convert_toTIME_EPOCH	(time32[ms]) -> binary
convert_toTIME_EPOCH_be	(time32[ms]) -> binary
convert_toTIMESTAMP_EPOCH	(timestamp[ms]) -> binary
convert_toTIMESTAMP_EPOCH_be	(timestamp[ms]) -> binary
convert_toDATE_EPOCH	(date64[ms]) -> binary
convert_toDATE_EPOCH_be	(date64[ms]) -> binary
months_between	(date64[ms], date64[ms]) -> double
	(timestamp[ms], timestamp[ms]) -> double
timestampdiffSecond	(timestamp[ms], timestamp[ms]) -> int32
timestampdiffMinute	(timestamp[ms], timestamp[ms]) -> int32
timestampdiffHour	(timestamp[ms], timestamp[ms]) -> int32
timestampdiffDay	(timestamp[ms], timestamp[ms]) -> int32
timestampdiffWeek	(timestamp[ms], timestamp[ms]) -> int32
timestampdiffMonth	(timestamp[ms], timestamp[ms]) -> int32
timestampdiffQuarter	(timestamp[ms], timestamp[ms]) -> int32
timestampdiffYear	(timestamp[ms], timestamp[ms]) -> int32
timestampaddSecond	(int32, timestamp[ms]) -> timestamp[ms]
	(int32, date64[ms]) -> date64[ms]
	(int64, timestamp[ms]) -> timestamp[ms]
	(int64, date64[ms]) -> date64[ms]
timestampaddMinute	(int32, timestamp[ms]) -> timestamp[ms]
	(int32, date64[ms]) -> date64[ms]
	(int64, timestamp[ms]) -> timestamp[ms]
	(int64, date64[ms]) -> date64[ms]
timestampaddHour	(int32, timestamp[ms]) -> timestamp[ms]
	(int32, date64[ms]) -> date64[ms]
	(int64, timestamp[ms]) -> timestamp[ms]
	(int64, date64[ms]) -> date64[ms]
timestampaddDay	(int32, timestamp[ms]) -> timestamp[ms]
	(int32, date64[ms]) -> date64[ms]
	(int64, timestamp[ms]) -> timestamp[ms]
	(int64, date64[ms]) -> date64[ms]
timestampaddWeek	(int32, timestamp[ms]) -> timestamp[ms]
	(int32, date64[ms]) -> date64[ms]
	(int64, timestamp[ms]) -> timestamp[ms]
	(int64, date64[ms]) -> date64[ms]
timestampaddMonth	(int32, timestamp[ms]) -> timestamp[ms]
	(int32, date64[ms]) -> date64[ms]
	(int64, timestamp[ms]) -> timestamp[ms]
	(int64, date64[ms]) -> date64[ms]
timestampaddQuarter	(int32, timestamp[ms]) -> timestamp[ms]
	(int32, date64[ms]) -> date64[ms]
	(int64, timestamp[ms]) -> timestamp[ms]
	(int64, date64[ms]) -> date64[ms]
timestampaddYear	(int32, timestamp[ms]) -> timestamp[ms]
	(int32, date64[ms]) -> date64[ms]
	(int64, timestamp[ms]) -> timestamp[ms]
	(int64, date64[ms]) -> date64[ms]
date_add	(date64[ms], int32) -> date64[ms]
	(timestamp[ms], int32) -> timestamp[ms]
	(date64[ms], int64) -> date64[ms]
	(timestamp[ms], int64) -> timestamp[ms]
	(int32, date64[ms]) -> date64[ms]
	(int32, timestamp[ms]) -> timestamp[ms]
	(int64, date64[ms]) -> date64[ms]
	(int64, timestamp[ms]) -> timestamp[ms]
date_sub	(date64[ms], int32) -> date64[ms]
	(timestamp[ms], int32) -> timestamp[ms]
	(date64[ms], int64) -> date64[ms]
	(timestamp[ms], int64) -> timestamp[ms]
date_diff	(date64[ms], int32) -> date64[ms]
	(timestamp[ms], int32) -> timestamp[ms]
	(date64[ms], int64) -> date64[ms]
	(timestamp[ms], int64) -> timestamp[ms]

String Transformations

to_date	(string, string) -> date64[ms]
	(string, string, int32) -> date64[ms]
sha	(string) -> string
	(binary) -> string
sha1	(string) -> string
	(binary) -> string
sha256	(string) -> string
	(binary) -> string
starts_with	(string, string) -> bool
ends_with	(string, string) -> bool
is_substr	(string, string) -> bool
locate	(string, string) -> int32
	(string, string, int32) -> int32
position	(string, string) -> int32
	(string, string, int32) -> int32
octet_length	(string) -> int32
	(binary) -> int32
bit_length	(string) -> int32
	(binary) -> int32
char_length	(string) -> int32
length	(string) -> int32
lengthUtf8	(binary) -> int32
reverse	(string) -> string
ltrim	(string) -> string
	(string, string) -> string
rtrim	(string) -> string
	(string, string) -> string
btrim	(string) -> string
	(string, string) -> string
ascii	(string) -> int32
base64	(binary) -> string
unbase64	(string) -> binary
upper	(string) -> string
lower	(string) -> string
initcap	(string) -> string
like	(string, string) -> bool
	(string, string, string) -> bool
ilike	(string, string) -> bool
substr	(string, int64, int64) -> string
	(string, int64) -> string
substring	(string, int64, int64) -> string
	(string, int64) -> string
lpad	(string, int32, string) -> string
	(string, int32) -> string
rpad	(string, int32, string) -> string
	(string, int32) -> string
concatOperator	(string, string) -> string
	(string, string, string) -> string
	(string, string, string, string) -> string
	(string, string, string, string, string) -> string
	(string, string, string, string, string, string) -> string
	(string, string, string, string, string, string, string) -> string
	(string, string, string, string, string, string, string, string) -> string
	(string, string, string, string, string, string, string, string, string) -> string
	(string, string, string, string, string, string, string, string, string, string) -> string
concat	(string, string) -> string
	(string, string, string) -> string
	(string, string, string, string) -> string
	(string, string, string, string, string) -> string
	(string, string, string, string, string, string) -> string
	(string, string, string, string, string, string, string) -> string
	(string, string, string, string, string, string, string, string) -> string
	(string, string, string, string, string, string, string, string, string) -> string
	(string, string, string, string, string, string, string, string, string, string) -> string
byte_substr	(binary, int32, int32) -> binary
bytesubstring	(binary, int32, int32) -> binary
convert_fromUTF8	(binary) -> string
convert_fromutf8	(binary) -> string
convert_replaceUTF8	(binary, string) -> string
convert_replaceutf8	(binary, string) -> string
convert_toUTF8	(string) -> binary
replace	(string, string, string) -> string
binary_string	(string) -> binary
left	(string, int32) -> string
right	(string, int32) -> string
split_part	(string, string, int32) -> string

Hash

hash	(int8) -> int32
	(int16) -> int32
	(int32) -> int32
	(int64) -> int32
	(uint8) -> int32
	(uint16) -> int32
	(uint32) -> int32
	(uint64) -> int32
	(float) -> int32
	(double) -> int32
	(decimal128(38, 0)) -> int32
	(date64[ms]) -> int32
	(timestamp[ms]) -> int32
	(time32[ms]) -> int32
	(date32[day]) -> int32
	(bool) -> int32
	(string) -> int32
	(binary) -> int32
hash32	(int8) -> int32
	(int16) -> int32
	(int32) -> int32
	(int64) -> int32
	(uint8) -> int32
	(uint16) -> int32
	(uint32) -> int32
	(uint64) -> int32
	(float) -> int32
	(double) -> int32
	(decimal128(38, 0)) -> int32
	(date64[ms]) -> int32
	(timestamp[ms]) -> int32
	(time32[ms]) -> int32
	(date32[day]) -> int32
	(bool) -> int32
	(string) -> int32
	(binary) -> int32
	(int8, int32) -> int32
	(int16, int32) -> int32
	(int32, int32) -> int32
	(int64, int32) -> int32
	(uint8, int32) -> int32
	(uint16, int32) -> int32
	(uint32, int32) -> int32
	(uint64, int32) -> int32
	(float, int32) -> int32
	(double, int32) -> int32
	(decimal128(38, 0), int32) -> int32
	(date64[ms], int32) -> int32
	(timestamp[ms], int32) -> int32
	(time32[ms], int32) -> int32
	(date32[day], int32) -> int32
	(bool, int32) -> int32
	(string, int32) -> int32
	(binary, int32) -> int32
hash32AsDouble	(int8) -> int32
	(int16) -> int32
	(int32) -> int32
	(int64) -> int32
	(uint8) -> int32
	(uint16) -> int32
	(uint32) -> int32
	(uint64) -> int32
	(float) -> int32
	(double) -> int32
	(decimal128(38, 0)) -> int32
	(date64[ms]) -> int32
	(timestamp[ms]) -> int32
	(time32[ms]) -> int32
	(date32[day]) -> int32
	(bool) -> int32
	(string) -> int32
	(binary) -> int32
	(int8, int32) -> int32
	(int16, int32) -> int32
	(int32, int32) -> int32
	(int64, int32) -> int32
	(uint8, int32) -> int32
	(uint16, int32) -> int32
	(uint32, int32) -> int32
	(uint64, int32) -> int32
	(float, int32) -> int32
	(double, int32) -> int32
	(decimal128(38, 0), int32) -> int32
	(date64[ms], int32) -> int32
	(timestamp[ms], int32) -> int32
	(time32[ms], int32) -> int32
	(date32[day], int32) -> int32
	(bool, int32) -> int32
	(string, int32) -> int32
	(binary, int32) -> int32
hash64	(int8) -> int64
	(int16) -> int64
	(int32) -> int64
	(int64) -> int64
	(uint8) -> int64
	(uint16) -> int64
	(uint32) -> int64
	(uint64) -> int64
	(float) -> int64
	(double) -> int64
	(decimal128(38, 0)) -> int64
	(date64[ms]) -> int64
	(timestamp[ms]) -> int64
	(time32[ms]) -> int64
	(date32[day]) -> int64
	(bool) -> int64
	(string) -> int64
	(binary) -> int64
	(int8, int64) -> int64
	(int16, int64) -> int64
	(int32, int64) -> int64
	(int64, int64) -> int64
	(uint8, int64) -> int64
	(uint16, int64) -> int64
	(uint32, int64) -> int64
	(uint64, int64) -> int64
	(float, int64) -> int64
	(double, int64) -> int64
	(decimal128(38, 0), int64) -> int64
	(date64[ms], int64) -> int64
	(timestamp[ms], int64) -> int64
	(time32[ms], int64) -> int64
	(date32[day], int64) -> int64
	(bool, int64) -> int64
	(string, int64) -> int64
	(binary, int64) -> int64
hash64AsDouble	(int8) -> int64
	(int16) -> int64
	(int32) -> int64
	(int64) -> int64
	(uint8) -> int64
	(uint16) -> int64
	(uint32) -> int64
	(uint64) -> int64
	(float) -> int64
	(double) -> int64
	(decimal128(38, 0)) -> int64
	(date64[ms]) -> int64
	(timestamp[ms]) -> int64
	(time32[ms]) -> int64
	(date32[day]) -> int64
	(bool) -> int64
	(string) -> int64
	(binary) -> int64
	(int8, int64) -> int64
	(int16, int64) -> int64
	(int32, int64) -> int64
	(int64, int64) -> int64
	(uint8, int64) -> int64
	(uint16, int64) -> int64
	(uint32, int64) -> int64
	(uint64, int64) -> int64
	(float, int64) -> int64
	(double, int64) -> int64
	(decimal128(38, 0), int64) -> int64
	(date64[ms], int64) -> int64
	(timestamp[ms], int64) -> int64
	(time32[ms], int64) -> int64
	(date32[day], int64) -> int64
	(bool, int64) -> int64
	(string, int64) -> int64
	(binary, int64) -> int64
hashSHA1	(int8) -> string
	(int16) -> string
	(int32) -> string
	(int64) -> string
	(uint8) -> string
	(uint16) -> string
	(uint32) -> string
	(uint64) -> string
	(float) -> string
	(double) -> string
	(decimal128(38, 0)) -> string
	(date64[ms]) -> string
	(timestamp[ms]) -> string
	(time32[ms]) -> string
	(date32[day]) -> string
	(bool) -> string
	(string) -> string
	(binary) -> string
hashSHA256	(int8) -> string
	(int16) -> string
	(int32) -> string
	(int64) -> string
	(uint8) -> string
	(uint16) -> string
	(uint32) -> string
	(uint64) -> string
	(float) -> string
	(double) -> string
	(decimal128(38, 0)) -> string
	(date64[ms]) -> string
	(timestamp[ms]) -> string
	(time32[ms]) -> string
	(date32[day]) -> string
	(bool) -> string
	(string) -> string
	(binary) -> string

Compute Functions Input / output and filesystems