チュートリアル: Streamlit を使用して Databricks アプリを開発する

このチュートリアルでは、 Databricks SQL Connector for PythonStreamlit を使用して Databricks アプリを構築する方法について説明します。 以下を実行するアプリを開発する方法について説明します。

  • Unity カタログ テーブルを読み取り、Streamlit インターフェイスに表示します。
  • データを編集し、テーブルに書き戻します。

手順 1: 特権を構成する

これらの例では、アプリがアプリの 承認を使用していることを前提としています。 アプリのサービス プリンシパルには、次が必要です。

  • Unity Catalog テーブルのSELECT 権限
  • Unity Catalog テーブルのMODIFY 権限
  • SQL ウェアハウスの CAN USE 権限

詳細については、 Unity カタログの権限リファレンスSQL Warehouse ACL に関するページを参照してください。

手順 2: 依存関係をインストールする

requirements.txt ファイルを作成し、次のパッケージを含めます。

databricks-sdk
databricks-sql-connector
streamlit
pandas

手順 3: アプリの実行を構成する

app.yaml ファイルを作成して、Azure Databricks Apps でのアプリの起動方法を定義します。

command: ['streamlit', 'run', 'app.py']

手順 4: Unity カタログ テーブルを読み取る

このコード例では、Unity カタログ テーブルからデータを読み取り、Streamlit を使用して表示する方法を示します。 次の目的を満たす app.py ファイルを作成します。

  • アプリ サービス プリンシパル認証を使用します。
  • SQL ウェアハウスの HTTP パスと Unity カタログ テーブル名の入力をユーザーに求めます。
  • 指定したテーブルに対して SELECT * クエリを実行します。
  • 結果を Streamlit st.dataframeで表示します。

app.py

import pandas as pd
import streamlit as st
from databricks import sql
from databricks.sdk.core import Config
import os

cfg = Config()

# Use app service principal authentication
def get_connection(http_path):
    server_hostname = cfg.host
    if server_hostname.startswith('https://'):
        server_hostname = server_hostname.replace('https://', '')
    elif server_hostname.startswith('http://'):
        server_hostname = server_hostname.replace('http://', '')
    return sql.connect(
        server_hostname=server_hostname,
        http_path=http_path,
        credentials_provider=lambda: cfg.authenticate,
        _use_arrow_native_complex_types=False,
    )

# Read data from a Unity Catalog table and return it as a pandas DataFrame
def read_table(table_name: str, conn) -> pd.DataFrame:
    with conn.cursor() as cursor:
        cursor.execute(f"SELECT * FROM {table_name}")
        return cursor.fetchall_arrow().to_pandas()

# Use Streamlit input fields to accept user input
http_path_input = st.text_input(
    "Enter your Databricks HTTP Path:", placeholder="/sql/1.0/warehouses/xxxxxx"
)
table_name = st.text_input(
    "Specify a Unity Catalog table name:", placeholder="catalog.schema.table"
)

# Display the result in a Streamlit DataFrame
if http_path_input and table_name:
    conn = get_connection(http_path_input)
    df = read_table(table_name, conn)
    st.dataframe(df)
else:
    st.warning("Provide both the warehouse path and a table name to load data.")

手順 5: Unity カタログ テーブルを編集する

このコード例では、Streamlit のデータ編集機能を使用して、Unity カタログ テーブルに対する変更の読み取り、編集、および書き込みをユーザーに許可します。 app.py ファイルに次の機能を追加します。

  • INSERT OVERWRITEを使用して、更新されたデータをテーブルに書き戻します。

app.py

import pandas as pd
import streamlit as st
from databricks import sql
from databricks.sdk.core import Config
import math

cfg = Config()

# Use app service principal authentication
def get_connection(http_path):
    server_hostname = cfg.host
    if server_hostname.startswith('https://'):
        server_hostname = server_hostname.replace('https://', '')
    elif server_hostname.startswith('http://'):
        server_hostname = server_hostname.replace('http://', '')
    return sql.connect(
        server_hostname=server_hostname,
        http_path=http_path,
        credentials_provider=lambda: cfg.authenticate,
        _use_arrow_native_complex_types=False,
    )

# Read data from a Unity Catalog table and return it as a pandas DataFrame
def read_table(table_name: str, conn) -> pd.DataFrame:
    with conn.cursor() as cursor:
        cursor.execute(f"SELECT * FROM {table_name}")
        return cursor.fetchall_arrow().to_pandas()

# Format values for SQL, handling NaN/None as NULL
def format_value(val):
    if val is None or (isinstance(val, float) and math.isnan(val)):
        return 'NULL'
    else:
        return repr(val)

# Use `INSERT OVERWRITE` to update existing rows and insert new ones
def insert_overwrite_table(table_name: str, df: pd.DataFrame, conn):
    progress = st.empty()
    with conn.cursor() as cursor:
        rows = list(df.itertuples(index=False))
        values = ",".join([f"({','.join(map(format_value, row))})" for row in rows])
        with progress:
            st.info("Calling Databricks SQL...")
        cursor.execute(f"INSERT OVERWRITE {table_name} VALUES {values}")
    progress.empty()
    st.success("Changes saved")

# Use Streamlit input fields to accept user input
http_path_input = st.text_input(
    "Enter your Databricks HTTP Path:", placeholder="/sql/1.0/warehouses/xxxxxx"
)
table_name = st.text_input(
    "Specify a Unity Catalog table name:", placeholder="catalog.schema.table"
)

# Display the result in a Streamlit DataFrame
if http_path_input and table_name:
    conn = get_connection(http_path_input)
    if conn:
        st.success("✅ Connected successfully!")
        original_df = read_table(table_name, conn)
        edited_df = st.data_editor(original_df, num_rows="dynamic", hide_index=True)
        df_diff = pd.concat([original_df, edited_df]).drop_duplicates(keep=False)
        if not df_diff.empty:
            st.warning(f"⚠️ You have {len(df_diff) // 2} unsaved changes")
            if st.button("Save changes"):
                insert_overwrite_table(table_name, edited_df, conn)
                st.rerun()
else:
    st.warning("Provide both the warehouse path and a table name to load data.")

次のステップ