Polars创建DataFrame时指定日期列失败，后续转换却成功

问题场景 #

拥有字典列表results，预期的Schema定义如下：

import polars as pl

results = ...
schema = {
    "id": pl.Int64,
    "title": pl.Utf8,
    "proprietors": pl.List(pl.Utf8),
    "registrationDate": pl.Date,
}

直接使用该Schema创建DataFrame时触发报错：

df = pl.DataFrame(results, schema=schema)

报错信息 #

ComputeError: could not append value: "2024-08-16T00:00:00" of type:
str to the builder; make sure that all rows have the same schema or
consider increasing infer_schema_length

it might also be that a value overflows the data-type's capacity

可行解决方案 #

先将日期列的类型临时改为pl.Utf8创建DataFrame，之后再转换为日期时间类型即可成功执行：

for col, dtype in schema.items():
    if col.endswith("Date"):
        schema[col] = pl.Utf8
df = pl.DataFrame(results, schema=schema)

for col, dtype in schema.items():
    if col.endswith("Date"):
        df = df.with_columns(pl.col(col).str.strptime(pl.Datetime))
# 执行成功，无报错

使用环境：polars 1.6.0，Python 3.12.4

内容的提问来源于stack exchange，提问作者tsorn