Charger des données à partir d'un DataFrame

Chargez le contenu d'un DataFrame pandas dans une table.

Exemple de code

Python

Avant d'essayer cet exemple, suivez les instructions de configuration pour Python du guide de démarrage rapide de BigQuery : Utiliser les bibliothèques clientes. Pour en savoir plus, consultez la documentation de référence de l'API BigQuery pour Python.

Pour vous authentifier auprès de BigQuery, configurez le service Identifiants par défaut de l'application. Pour en savoir plus, consultez la page Configurer l'authentification pour les bibliothèques clientes.

import datetime

from google.cloud import bigquery
import pandas
import pytz

# Construct a BigQuery client object.
client = bigquery.Client()

# TODO(developer): Set table_id to the ID of the table to create.
# table_id = "your-project.your_dataset.your_table_name"

records = [
    {
        "title": "The Meaning of Life",
        "release_year": 1983,
        "length_minutes": 112.5,
        "release_date": pytz.timezone("Europe/Paris")
        .localize(datetime.datetime(1983, 5, 9, 13, 0, 0))
        .astimezone(pytz.utc),
        # Assume UTC timezone when a datetime object contains no timezone.
        "dvd_release": datetime.datetime(2002, 1, 22, 7, 0, 0),
    },
    {
        "title": "Monty Python and the Holy Grail",
        "release_year": 1975,
        "length_minutes": 91.5,
        "release_date": pytz.timezone("Europe/London")
        .localize(datetime.datetime(1975, 4, 9, 23, 59, 2))
        .astimezone(pytz.utc),
        "dvd_release": datetime.datetime(2002, 7, 16, 9, 0, 0),
    },
    {
        "title": "Life of Brian",
        "release_year": 1979,
        "length_minutes": 94.25,
        "release_date": pytz.timezone("America/New_York")
        .localize(datetime.datetime(1979, 8, 17, 23, 59, 5))
        .astimezone(pytz.utc),
        "dvd_release": datetime.datetime(2008, 1, 14, 8, 0, 0),
    },
    {
        "title": "And Now for Something Completely Different",
        "release_year": 1971,
        "length_minutes": 88.0,
        "release_date": pytz.timezone("Europe/London")
        .localize(datetime.datetime(1971, 9, 28, 23, 59, 7))
        .astimezone(pytz.utc),
        "dvd_release": datetime.datetime(2003, 10, 22, 10, 0, 0),
    },
]
dataframe = pandas.DataFrame(
    records,
    # In the loaded table, the column order reflects the order of the
    # columns in the DataFrame.
    columns=[
        "title",
        "release_year",
        "length_minutes",
        "release_date",
        "dvd_release",
    ],
    # Optionally, set a named index, which can also be written to the
    # BigQuery table.
    index=pandas.Index(
        ["Q24980", "Q25043", "Q24953", "Q16403"], name="wikidata_id"
    ),
)
job_config = bigquery.LoadJobConfig(
    # Specify a (partial) schema. All columns are always written to the
    # table. The schema is used to assist in data type definitions.
    schema=[
        # Specify the type of columns whose type cannot be auto-detected. For
        # example the "title" column uses pandas dtype "object", so its
        # data type is ambiguous.
        bigquery.SchemaField("title", bigquery.enums.SqlTypeNames.STRING),
        # Indexes are written if included in the schema by name.
        bigquery.SchemaField("wikidata_id", bigquery.enums.SqlTypeNames.STRING),
    ],
    # Optionally, set the write disposition. BigQuery appends loaded rows
    # to an existing table by default, but with WRITE_TRUNCATE write
    # disposition it replaces the table with the loaded data.
    write_disposition="WRITE_TRUNCATE",
)

job = client.load_table_from_dataframe(
    dataframe, table_id, job_config=job_config
)  # Make an API request.
job.result()  # Wait for the job to complete.

table = client.get_table(table_id)  # Make an API request.
print(
    "Loaded {} rows and {} columns to {}".format(
        table.num_rows, len(table.schema), table_id
    )
)

Étapes suivantes

Pour rechercher et filtrer des exemples de code pour d'autres produits Google Cloud, consultez l'explorateur d'exemples Google Cloud.