Skip to main content

Linking rows of pandas dataframes

Project description

# pandas-linker

`pandas-linker` runs comparison windows over different sortings of a pandas DataFrame and links the rows via assigned UUIDs. This library does not actually do any duplicate detection. Instead it provides a harness to run your own comparison functions on your data.

This library is meant for datasets of a size where comparing every row with every other is undesirable. Instead you can decide on a sorting order of the DataFrame and only compare every row with every other inside a sliding window.

## Install

```
pip install pandas-linker
```

## Example

Let's say you have a DataFrame like this:

[ix] | name | country
------|------|--------
0 | Pete | Spain
1 | Mary | USA
2 | Bart | US
3 | Mary | US

and you want to detect similar rows and mark them as such. Here's how to do that:

```python
from pandas_linker import get_linker


def compare_rows(a, b):
''' Example function that decides if two rows represent same entity.'''
return a['name'] in b['name'] or b['name'] in a['name']

# df is a pandas.DataFrame with a unique index

with get_linker(df, field='uid') as linker:

print('Comparing in 10 row window sorted by name')
linker(sort_cols=['name'], window_size=10, cmp=compare_rows)

print('Comparing in 15 row window sorted by country')
linker(sort_cols=['country'], window_size=15, cmp=compare_rows)

```

After running the linker the process is complete

[ix] | name | country | uid
------|------|---------|----
0 | Pete | Spain | 7509781940fc471cad5dc32944652d70
1 | Mary | USA | 8f8dccd91568472daf740e9160349d6c
2 | Bart | US | 12b55fbe80f64d378193acd727b0e051
3 | Mary | US | 8f8dccd91568472daf740e9160349d6c

Note that both "Mary" rows in the DataFrame have been identified as representing
the same entity and were assigned the same UUID.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Filename, size & hash SHA256 hash help File type Python version Upload date
pandas_linker-0.0.2-py2.py3-none-any.whl (6.0 kB) Copy SHA256 hash SHA256 Wheel py2.py3 Nov 24, 2016

Supported by

Elastic Elastic Search Pingdom Pingdom Monitoring Google Google BigQuery Sentry Sentry Error logging AWS AWS Cloud computing DataDog DataDog Monitoring Fastly Fastly CDN SignalFx SignalFx Supporter DigiCert DigiCert EV certificate StatusPage StatusPage Status page