Transforming your dataset

On this page we’ll guide you through some of the most common operations used when doing data analysis. This is only a small subsection of what is possible in Polars, for more information please visit the Documentation.

For the example we will use the Danbooru dataset. Danbooru is a large-scale anime image dataset with over 5 million images contributed and annotated in detail by an enthusiast community. Image tags cover aspects like characters, scenes, copyrights, artists, etc with an average of 30 tags per image.

Reading

df = pl.read_parquet("hf://datasets/KBlueLeaf/danbooru2023-metadata-database/parquet/post.parquet")
df.head(3)
┌─────┬────────────┬────────────┬────────────┬───┬────────────┬────────────┬───────────┬───────────┐
│ id  ┆ created_at ┆ uploader_i ┆ source     ┆ … ┆ file_url   ┆ large_file ┆ preview_f ┆ updated_a │
│ --- ┆ ---        ┆ d          ┆ ---        ┆   ┆ ---        ┆ _url       ┆ ile_url   ┆ t         │
│ i64 ┆ str        ┆ ---        ┆ str        ┆   ┆ str        ┆ ---        ┆ ---       ┆ ---       │
│     ┆            ┆ i64        ┆            ┆   ┆            ┆ str        ┆ str       ┆ str       │
╞═════╪════════════╪════════════╪════════════╪═══╪════════════╪════════════╪═══════════╪═══════════╡
│ 1   ┆ 2005-05-23 ┆ 1          ┆ http://www ┆ … ┆ https://cd ┆ https://cd ┆ https://c ┆ 2024-02-2 │
│     ┆ T23:35:31. ┆            ┆ .biwa.ne.j ┆   ┆ n.donmai.u ┆ n.donmai.u ┆ dn.donmai ┆ 3T23:16:2 │
│     ┆ 000-04:00  ┆            ┆ p/~kyogoku ┆   ┆ s/original ┆ s/original ┆ .us/180x1 ┆ 3.261-05: │
│     ┆            ┆            ┆ …          ┆   ┆ …          ┆ …          ┆ 80/…      ┆ 00        │
│ 2   ┆ 2005-05-23 ┆ 1          ┆ http://pag ┆ … ┆ https://cd ┆ https://cd ┆ https://c ┆ 2023-05-2 │
│     ┆ T23:37:30. ┆            ┆ e.freett.c ┆   ┆ n.donmai.u ┆ n.donmai.u ┆ dn.donmai ┆ 7T21:04:5 │
│     ┆ 000-04:00  ┆            ┆ om/tyuyan/ ┆   ┆ s/original ┆ s/original ┆ .us/180x1 ┆ 0.554-04: │
│     ┆            ┆            ┆ …          ┆   ┆ …          ┆ …          ┆ 80/…      ┆ 00        │
│ 3   ┆ 2005-05-23 ┆ 1          ┆ http://pag ┆ … ┆ https://cd ┆ https://cd ┆ https://c ┆ 2023-12-0 │
│     ┆ T23:38:05. ┆            ┆ e.freett.c ┆   ┆ n.donmai.u ┆ n.donmai.u ┆ dn.donmai ┆ 5T08:21:0 │
│     ┆ 000-04:00  ┆            ┆ om/tyuyan/ ┆   ┆ s/original ┆ s/original ┆ .us/180x1 ┆ 8.958-05: │
│     ┆            ┆            ┆ …          ┆   ┆ …          ┆ …          ┆ 80/…      ┆ 00        │
└─────┴────────────┴────────────┴────────────┴───┴────────────┴────────────┴───────────┴───────────┘

Selecting columns

The dataset contains quite a number of columns in order to make the output more managable in the guide, we will select a number of columns:

df = df.select(["id","uploader_id","source","file_url","score"])
┌─────┬─────────────┬─────────────────────────────────┬─────────────────────────────────┬───────┐
│ id  ┆ uploader_id ┆ source                          ┆ file_url                        ┆ score │
│ --- ┆ ---         ┆ ---                             ┆ ---                             ┆ ---   │
│ i64 ┆ i64         ┆ str                             ┆ str                             ┆ i64   │
╞═════╪═════════════╪═════════════════════════════════╪═════════════════════════════════╪═══════╡
│ 1   ┆ 1           ┆ http://www.biwa.ne.jp/~kyogoku… ┆ https://cdn.donmai.us/original… ┆ 685   │
│ 2   ┆ 1           ┆ http://page.freett.com/tyuyan/… ┆ https://cdn.donmai.us/original… ┆ 13    │
│ 3   ┆ 1           ┆ http://page.freett.com/tyuyan/… ┆ https://cdn.donmai.us/original… ┆ 31    │
└─────┴─────────────┴─────────────────────────────────┴─────────────────────────────────┴───────┘

Filtering

We can filter the dataset using .filter(..) within a filter you can put complex expressions but let us start of easy. To filter based on the source of the image use the following:

df.filter(pl.col("source").str.contains("archive.org"))

You can combine multiple filters with & or ’|’ operators:

df.filter(pl.col("source").str.contains("archive.org") | (pl.col("uploader_id") > 1000))

Transforming

In order to add new columns to the dataset use with_columns. In the example below we look at the source column to determine if the traffic is encrypted and add a new column called is_https_url with the alias method. The entire statement within with_columns is called an expression. To read more about expressions and how to use them in the Polars User Guide

df.with_columns(pl.col("source").str.starts_with("https").alias("is_https_url"))

Aggregation & Sorting

In order to aggregate data together you can use the group_by, agg and sort methods:

df.group_by("uploader_id").agg(pl.len().alias("number_of_uploads")).sort("number_of_uploads",descending=True)
┌─────────────┬───────────────────┐
│ uploader_id ┆ number_of_uploads │
│ ---         ┆ ---               │
│ i64         ┆ u32               │
╞═════════════╪═══════════════════╡
│ 430030      ┆ 221689            │
│ 49091       ┆ 148006            │
│ 1           ┆ 136251            │
│ 30072       ┆ 114557            │
│ 483749      ┆ 95415             │
│ …           ┆ …                 │
│ 1028831     ┆ 1                 │
│ 1111802     ┆ 1                 │
│ 911109      ┆ 1                 │
│ 585897      ┆ 1                 │
│ 348393      ┆ 1                 │
└─────────────┴───────────────────┘
< > Update on GitHub