How do I batch upsert data into Google Cloud Spanner using the Python client library?

Question

I would like to upsert the contents of a pandas dataframe into a table in a Google Cloud Spanner database. The documentation here recommends using the insert_or_update() method of the batch object.

If the batch object is created by running this

from google.cloud import spanner_v1
client = spanner_v1.Client()
batch = client.batch()

Then this object does not have that method available. Running dir(client) gives me these results

['SCOPE', 
'_SET_PROJECT', 
'__class__', 
'__delattr__', 
'__dict__', 
'__dir__', 
'__doc__', 
'__eq__', 
'__format__', 
'__ge__', 
'__getattribute__', 
'__getstate__', 
'__gt__', 
'__hash__', 
'__init__', 
'__init_subclass__', 
'__le__', 
'__lt__', 
'__module__', 
'__ne__', 
'__new__', 
'__reduce__', 
'__reduce_ex__', 
'__repr__', 
'__setattr__', 
'__sizeof__', 
'__str__', 
'__subclasshook__', 
'__weakref__', 
'_credentials', 
'_database_admin_api', 
'_determine_default', 
'_http', 
'_http_internal', 
'_instance_admin_api', 
'_item_to_instance', 
'copy', 
'credentials', 
'database_admin_api', 
'from_service_account_json', 
'instance', 
'instance_admin_api', 
'list_instance_configs', 
'list_instances', 
'project', 
'project_name', 
'user_agent']

How do I do batch upsert in Spanner?

Answer 1

The snippets has an example of batch insert. I checked that the batch object created in the snippet also has an insert_or_update field.

https://github.com/GoogleCloudPlatform/python-docs-samples/blob/master/spanner/cloud-client/snippets.py#L72

[' class ', ' delattr ', ' dict ', ' doc ', ' enter ', ' exit ', ' format ', ' getattribute ', ' hash ', ' init ', ' module ', ' new ', ' reduce ', ' reduce_ex ', ' repr ', ' setattr ', ' sizeof ', ' str ', ' subclasshook ', ' weakref ', '_check_state', '_mutations', '_session', 'commit', 'committed', 'delete', 'insert', 'insert_or_update', 'replace', 'update']

Can you try that out?

Answer 2

If you have a pandas dataframe, here a random 5 x 3 with columns a,b,c, you can transform the dataframe to column names and the rows and batch insert.

import pandas as pd
import numpy as np

df = pd.DataFrame(np.random.randint(low=0, high=10, size=(5, 3)),
                  columns=['a', 'b', 'c'])

You can insert this into Google Cloud Spanner by extracting the columns and values from df and batch inserting.

from google.cloud import spanner

spanner_client = spanner.Client()
instance = spanner_client.instance(instance_id)
database = instance.database(database_id)

columns = df.columns
values = df.values.tolist()

with database.batch() as batch:
    batch.insert(
        table='table',
        columns=columns
        values=values
    )