3.4.5 Example Text Data

As of MySQL 8.1.0, HeatWave AutoML supports text data types. To create a sample text data set, use the fetch_20newsgroups data set from scikit-learn. This also uses the pandas Python library.

$> from sklearn.datasets import fetch_20newsgroups
$> import pandas as pd

$> categories = ['alt.atheism', 'talk.religion.misc', 'comp.graphics', 'sci.space']

$> newsgroups_train = fetch_20newsgroups(subset='train', remove=('headers', 'footers', 'quotes'), categories=categories)
$> df = pd.DataFrame([newsgroups_train.data, newsgroups_train.target.tolist()]).T
$> df.columns = ['text', 'target']
$> targets = pd.DataFrame( newsgroups_train.target_names)
$> targets.columns=['category']
$> out = pd.merge(df, targets, left_on='target', right_index=True).drop('target', axis=1)
$> out = out[(out.text != '')] #remove empty strings
$> out.to_csv('20newsgroups_train.csv', index=False)

$> newsgroups_test = fetch_20newsgroups(subset='test', remove=('headers', 'footers', 'quotes'), categories=categories)
$> df = pd.DataFrame([newsgroups_test.data, newsgroups_test.target.tolist()]).T
$> df.columns = ['text', 'target']
$> targets = pd.DataFrame( newsgroups_test.target_names)
$> targets.columns=['category']
$> out = pd.merge(df, targets, left_on='target', right_index=True).drop('target', axis=1)
$> out = out[(out.text != '')] #remove empty strings
$> out.to_csv('20newsgroups_test.csv', index=False)

Then load the csv files into MySQL:

mysql> DROP TABLE IF EXISTS `20newsgroups_train`;
mysql> DROP TABLE IF EXISTS `20newsgroups_test`;
mysql> CREATE TABLE `20newsgroups_train` (`text` LONGTEXT DEFAULT NULL, `target` VARCHAR(255) DEFAULT NULL);
mysql> CREATE TABLE `20newsgroups_test` LIKE `20newsgroups_train`;

mysql-js> util.importTable("20newsgroups_train.csv",{table: "20newsgroups_train", dialect: "csv-unix", skipRows:1})
mysql-js> util.importTable("20newsgroups_test.csv",{table: "20newsgroups_test", dialect: "csv-unix", skipRows:1})