使用Pandas获取Eurostat数据的两类工具问题求助

Let's tackle your two tool issues one by one, with clear code examples and explanations tailored to your use case.

1. Fixing Data Filtering with pandas_datareader #

The pandas_datareader Eurostat integration does support filtering—you just need to pass the right parameters to narrow down to TOTAL or NEW dwelling types. Here's how:

First, confirm the dimension codes for the prc_hpi_a dataset (you can look these up directly on the Eurostat website's dataset page). For housing type, the relevant dimension is indic_bt, with codes:

• TOTAL: All dwellings
• NEW: New dwellings
• EXISTING: Existing dwellings (the default you're seeing)

Use the params argument in pandas_datareader.data.get_data_eurostat() to specify your filter:

import pandas_datareader.data as web

# Fetch filtered data for TOTAL and NEW dwellings
df = web.get_data_eurostat('prc_hpi_a', 
                           params={'indic_bt': ['TOTAL', 'NEW']},
                           start='2010-01-01')

# Reset index to make dimensions more readable
df = df.reset_index()
print(df.head())

If you're unsure about dimension names, first fetch the dataset's metadata to list all available filters:

metadata = web.get_data_eurostat('prc_hpi_a', metadata=True)
print(metadata['dimensions'])

2. Resolving pandasdmx Issues #

The pandasdmx library has undergone breaking changes in recent versions, which is why old tutorials aren't working. Let's fix both your problems:

a. Fixing ImportError: cannot import name 'client' #

The old client module was renamed to Request in newer versions. Update your import line to:

from pandasdmx import Request

Then initialize the Eurostat client like this:

estat = Request('ESTAT')

b. Fixing AttributeError: 'DataMessage' object has no attribute 'data' #

Some Eurostat datasets store data in .data_set instead of .data within the DataMessage object. Here's a reliable way to extract and convert to a DataFrame:

# Fetch the prc_hicp_midx dataset
resp = estat.data('prc_hicp_midx')

# Handle different data storage attributes
if hasattr(resp, 'data_set'):
    data = resp.data_set
else:
    data = resp.data

# Convert to Pandas DataFrame
df = resp.write(data)
print(df.head())

Alternatively, use the newer .to_pandas() method which automatically handles both cases:

df = resp.to_pandas()

For targeted filtering, specify keys when fetching data:

# Example: Filter prc_hicp_midx by country (DE, FR) and main indicator
resp = estat.data('prc_hicp_midx', key={'geo': ['DE', 'FR'], 'indic': 'CP00'})
df = resp.to_pandas()

Additional Tips #

• Always cross-reference the Eurostat dataset page for dimension codes and structure—this will speed up your filtering workflow.
• In Azure Notebooks, ensure you have the latest library versions installed:

  !pip install --upgrade pandas_datareader pandasdmx
  

内容的提问来源于stack exchange，提问作者SPS