dask_dataframe_groupby_aggregation_with_local_client.py

1import dask.dataframe as dd
2import pandas as pd
3from dask.distributed import Client
4
5# 1. Setup a local cluster and client
6# This provides a dashboard at http://localhost:8787
7client = Client()
8
9# 2. Create a dummy pandas DataFrame
10df = pd.DataFrame({
11    "name": ["Alice", "Bob", "Charlie", "Dan", "Edith", "Frank"],
12    "age": [25, 30, 35, 40, 45, 50],
13    "amount": [100, 200, 300, 400, 500, 600]
14})
15
16# 3. Convert to a Dask DataFrame with 2 partitions
17ddf = dd.from_pandas(df, npartitions=2)
18
19# 4. Perform a calculation (lazy evaluation)
20# Dask doesn't compute the result immediately
21result_task = ddf.groupby("name").amount.mean()
22
23# 5. Compute the result
24# This triggers the actual execution
25result = result_task.compute()
26
27print(result)
28
29# 6. Close the client
30client.close()