DataFrame.from_records() Function | Teradata Package for Python - DataFrame.from_records Function - Teradata Package for Python

Use the DataFrame.from_records() function to create a DataFrame from a list of lists/tuples/dictionaries/numpy arrays.

Required Parameter

data

Specifies the iterator of data or the list of lists/tuples/dictionaries/numpy arrays to be converted to teradataml DataFrame.

Nested lists or tuples or dictionaries are not supported.

Optional Parameters

columns

Specifies the column names for the DataFrame.

If the data is a list of lists/tuples/numpy arrays and this argument is not specified, column names will be auto-generated as 'col_0', 'col_1', etc.

kwargs

Specifies the columns to be excluded from the DataFrame.

coerce_float

Specifies whether to convert values of non-string, non-numeric objects (like decimal.Decimal) to floating point, useful for SQL result sets.

nrows

Specifies the number of rows to be read from the data if the data is iterator.

persist

Specifies whether to persist the DataFrame.

Default value: false

Example setup

>>> from teradataml import DataFrame

Example 1: Create a teradataml DataFrame from a list of lists

>>> df = DataFrame.from_records([['Alice', 1], ['Bob', 2]], columns=['name', 'age'])
>>> df

    name age
0    Bob   2
1  Alice   1

Example 2: Create a teradataml DataFrame from a list of tuples

>>> df = DataFrame.from_records([('Alice', 1), ('Bob', 3)], columns=['name', 'age'])
>>> df

    name age
0    Bob   3
1  Alice   1

Example 3: Create a teradataml DataFrame from a list of dictionaries

>>> df = DataFrame.from_records([{'name': 'Alice', 'age': 4}, {'name': 'Bob', 'age': 2}])
>>> df

    name age
0    Bob   2
1  Alice   4

Example 4: Create a teradataml DataFrame from a list where columns are not explicitly defined

>>> df = DataFrame.from_records([['Alice', 1], ['Bob', 2]])
>>> df

    col_0 col_1
0     Bob     2
1   Alice     1

Example 5: Create a teradataml DataFrame from a list by excluding 'grade' column

>>> df = DataFrame.from_records([['Alice', 1, 'A'], ['Bob', 2, 'B']],
... columns=['name', 'age', 'grade'],
... exclude=['grade'])
>>> df

    name   age
0    Bob     2
1  Alice     1

Example 6: Create a teradataml DataFrame from a list of lists with "coerce_float" set to False

>>> df = DataFrame.from_records([[1, Decimal('2.5')], [3, Decimal('4.0')]],
... columns=['col1', 'col2'], coerce_float=False)
>>> df

    col1   col2
0     3     4.0
1     1     2.5

>>> df.tdtypes

col1 BIGINT()
col2 VARCHAR(length=1024, charset='UNICODE')

Example 7: Create a teradataml DataFrame from a list of lists with "coerce_float" set to True

>>> from decimal import Decimal
>>> df = DataFrame.from_records([[1, Decimal('2.5')], [3, Decimal('4.0')]],
... columns=['col1', 'col2'], coerce_float=True)
>>> df

    col1 col2
0      3  4.0
1      1  2.5

>>> df.tdtypes

col1 BIGINT()
col2 FLOAT()

Example 8: Create a teradataml DataFrame from an iterator with 'nrows' set to 2

>>> def data_gen():
... yield ['Alice', 1]
... yield ['Bob', 2]
... yield ['Charlie', 3]
>>> df = DataFrame.from_records(data_gen(), columns=['name', 'age'], nrows=2)
>>> df

name age
0 Bob 2
1 Alice 1

Example 9: Persist the data from records into a table which can be used across sessions

 >>> df = DataFrame.from_records([['Alice', 1], ['Bob', 2]], columns=['name', 'age'], persist=True)
>>> df

            name  age
        0    Bob    2
        1  Alice    1

Get the database object name.

>>> df.db_object_name

'"ml__from_pandas_1761719973588452"'

Create the teradataml DataFrame using the database object name in different session.

>>> DataFrame('"ml__from_pandas_1761719973588452"')

            name  age
        0    Bob    2
        1  Alice    1