Types Overview

In Python, 5 is of type int while "hello" is of type str. Python is a dynamically-typed language, meaning that a function like:

>>> def add_x_and_y(x, y):
...     return x + y

can be called on any two objects which can be added, like numbers, strings, or numpy arrays.

Types are very important in Hail, because the fields of Table and MatrixTable objects have data types.

Hail’s primitive data types for boolean, numeric and string objects are:

  • tstr - Text string.

  • tbool - Boolean (True or False) value.

  • tint32 - 32-bit integer.

  • tint64 - 64-bit integer.

  • tfloat32 - 32-bit floating point number.

  • tfloat64 - 64-bit floating point number.

Hail’s container types are:

  • tarray - Ordered collection of homogenous objects.

  • tndarray - Ordered n-dimensional arrays of homogenous objects.

  • tset - Unordered collection of distinct homogenous objects.

  • tdict - Key-value map. Keys and values are both homogenous.

  • ttuple - Tuple of heterogeneous values.

  • tstruct - Structure containing named fields, each with its own type.

Hail also has a few genetics-specific types:

  • tcall - Genotype calls.

  • tlocus - Genomic locus, parameterized by reference genome.

When to work with types

In general, you won’t need to mention types explicitly. Hail will automatically impute the type of your data.

There are a few situations where you may want to specify types explicitly:

  • To specify column types in import_table() if the imputed types do not match what you want.

  • When converting a Python value to a Hail expression with literal(), if you don’t wish to rely on the imputed type.

  • When using missing types via the null() constructor.

Viewing an object’s type

Hail objects have a dtype field that will print their type.

>>> hl.int32(3).dtype

Entering just the object will also give you some type information.

>>> hl.int32(3)
<Int32Expression of type int32>

We can see that hl.int32(3) is of type tint32, but what does Expression mean? Each data type in Hail is represented by its own Expression class. Data of type tint32 is represented by an Int32Expression. Data of type tstruct is represented by a StructExpression.

If you examine the type of a container object, such as a struct, you’ll notice that the struct expression’s type also contains the subtypes of the nested fields.

>>> hl.struct(name='Hail', dob=2015)
<StructExpression of type struct{name: str, dob: int32}>
>>> hl.struct(name='Hail', dob=2015).dtype
dtype('struct{name: str, dob: int32}')

Container Types

Hail’s container types for arrays, sets, dicts, and tuples require homogenous collections, meaning that all values in the collection must be of the same type. In contrast, Python allows mixed collections, e.g. ['1', 2, 3.0] is a valid Python list. A Hail array could not contain both tstr and tint32 objects. Likewise, the dict {'a': 1, 2: 'b'} is a valid Python dictionary, but a Hail dictionary cannot contain keys of different types. An example of a valid dictionary in Hail is {'a': 1, 'b': 2}, where the keys are all strings and the values are all integers. The type of this dictionary would be dict<str, int32>.


Hail’s tstruct type is used to compose types together to form nested structures. Structs can contain any combination of types. The tstruct is an ordered mapping from field name to field type. Each field name must be unique. So a struct hl.struct(name='Hail', dob=2015) has type dtype('struct{name: str, dob: int32}') and contains a mapping from name to a string field and from dob to integer fields.

Structs are very common in Hail. Consider:

>>> new_table = table1.annotate(table2_fields = table2[table1.key])

This snippet adds a field to table1 called table2_fields. In the new table, table2_fields will be a struct containing all the nested fields from table2.