- Load the dataset from example collection.
load_example_data('teradataml','bank_marketing') bank_df = DataFrame("bank_marketing") bank_dfOut:
bank_df age job marital education default_value balance housing loan contact day_of_month month_of_year duration campaign pdays previous poutcome deposit 45 admin. married secondary no 149 yes no unknown 23 may 893 3 -1 0 unknown yes 45 blue-collar divorced primary no 594 yes no unknown 29 may 833 2 -1 0 unknown yes 45 technician single secondary no 410 yes no unknown 30 may 891 4 -1 0 unknown yes 45 management divorced secondary no 644 yes no unknown 4 jun 633 1 -1 0 unknown yes 45 management married tertiary no 655 no no unknown 20 jun 693 3 -1 0 unknown yes 45 technician married secondary no 879 no no cellular 7 jul 621 2 -1 0 unknown yes 45 blue-collar divorced primary no 844 no no unknown 5 jun 1018 3 -1 0 unknown yes 45 unemployed divorced secondary no 3354 yes no unknown 29 may 746 1 -1 0 unknown yes 45 blue-collar divorced primary no -311 yes no unknown 23 may 1030 1 -1 0 unknown yes 45 entrepreneur divorced tertiary no -395 yes no unknown 13 may 470 1 -1 0 unknown yes
- Add a partition column to create a single model with all data
new_partition_columns = ["partition_column_1"] bank_df = bank_df.assign(**{new_partition_columns[0]: 1001}) - Create test and train data using Teradata's in-database functions.
bank_df_sample = bank_df.sample(frac = [0.8, 0.2]) bank_df_train= bank_df_sample[bank_df_sample['sampleid'] == 1].drop('sampleid', axis=1) bank_df_test = bank_df_sample[bank_df_sample['sampleid'] == 2].drop('sampleid', axis=1) - Train the data shape.
bank_df_train.shape
Out:
(8930, 18)
- Test the data shape.
bank_df_test.shape
Out:
(2232, 18)