Use str.extract to extract multiple matches in polars

Question

df = pl.DataFrame(
    {
        "a": [
            "name=John, name=Billy",
            "name=Jeff",
            "name=Taylor",
        ]
    }
)

df.select(
    pl.col("a").str.extract(r"name=(\w+)", 1),
)

I'll get a Series with John, Jeff, Taylor. I'm wondering if there's a way to extract_all(?) and also get Billy? I realize this changes the dimensionality of the resulting series but was just wondering if this method is available somehow.

Answer 1

Actually, there is a str.extract_all method.

df.select(
    pl.col("a")
    .str.extract_all(r"name=\w+")
    .explode()
    .str.extract(r"name=(\w+)")
    .alias("names")
)

shape: (4, 1)
┌────────┐
│ names  │
│ ---    │
│ str    │
╞════════╡
│ John   │
│ Billy  │
│ Jeff   │
│ Taylor │
└────────┘