Query a public dataset with the BigQuery Python client library

Open the Cloud Shell Editor

1. In Cloud Shell, create a new Python project and file:

    mkdir  
   bigquery-python-quickstart  
    \ 
    && 
   touch  
    \ 
     
   bigquery-python-quickstart/app.py 
   

   This command creates a Python project that's named bigquery-python-quickstart and a file that's named app.py .

2. Open the Cloud Shell Editor:

    cloudshell  
   workspace  
   bigquery-python-quickstart 
   

To learn how to prepare files for queries, click Next.

Prepare files for queries

1. To open a terminal in the Cloud Shell Editor, click Open Terminal.

2. Open your project directory:

     cd 
     
   bigquery-python-quickstart 
   

3. Install the BigQuery client library for Python:

    pip  
   install  
   --upgrade  
   google-cloud-bigquery 
   

   The output is similar to the following. Several lines are omitted to simplify the output.

    Installing collected packages: google-cloud-bigquery
   ...
   Successfully installed google-cloud-bigquery-3.9.0
   ... 
   

To learn how to query a public dataset in BigQuery, click Next.

Query a public dataset in BigQuery

1. Click Open Editor.

2. In the Explorerpane, locate your BIGQUERY-PYTHON-QUICKSTART project.

3. Click the app.py file to open it.

4. To create a query against the bigquery-public-data.stackoverflow dataset that returns the top 10 most viewed Stack Overflow pages and their view counts, copy the following code into the app.py file:

     from 
     
    google.cloud 
     
    import 
     bigquery 
    
    def 
     
    query_stackoverflow 
    () 
    - 
   > None 
    : 
    client 
    = 
     bigquery 
    
    . 
     Client 
    
    () 
    results 
    = 
    client 
    . 
     query_and_wait 
    
    ( 
     
    """ 
    SELECT 
    CONCAT( 
    'https://stackoverflow.com/questions/', 
    CAST(id as STRING)) as url, 
    view_count 
    FROM `bigquery-public-data.stackoverflow.posts_questions` 
    WHERE tags like '%google-bigquery%' 
    ORDER BY view_count DESC 
    LIMIT 10""" 
    ) 
    # Waits for job to complete. 
    for 
    row 
    in 
    results 
    : 
    print 
    ( 
    " 
    {} 
    : 
    {} 
    views" 
    . 
    format 
    ( 
    row 
    . 
    url 
    , 
    row 
    . 
    view_count 
    )) 
    if 
    __name__ 
    == 
    "__main__" 
    : 
    query_stackoverflow 
    ()