Lexically-ordered Categorical columns created in StringCache context lose their name following a sort

[alexander-beedie]

Polars version checks

• I have checked that this issue has not already been reported.

• I have confirmed this bug exists on the latest version of Polars.

Issue description

We're investigating a number of Categorical-related gremlins at the moment.
Here's the first, suitably minimised (against latest polars build 0.15.11).

Reproducible example

Setup:

import polars as pl

with pl.StringCache():
    # create basic categorical
    df1 = pl.DataFrame( {'key':['bbb','aaa','ccc'] } ).with_column(
        pl.col('key').cast(pl.Categorical)
    )
    # create lexically-ordered categorical
    df2 = pl.DataFrame( {'key':['bbb','aaa','ccc'] } ).with_column(
        pl.col('key').cast(pl.Categorical).cat.set_ordering('lexical') 
    )

# create lexically-ordered categorical *outside* of StringCache context
df3 = pl.DataFrame( {'key':['bbb','aaa','ccc'] } ).with_column(
    pl.col('key').cast(pl.Categorical).cat.set_ordering('lexical') 
)

Issue:

df1.sort( by=pl.col('key') )
# ┌─────┐
# │ key │
# │ --- │
# │ cat │
# ╞═════╡
# │ bbb │
# │ aaa │
# │ ccc │
# └─────┘

df2.sort( by=pl.col('key') )
# ┌─────┐
# │     │  << column name vanished
# │ --- │
# │ cat │
# ╞═════╡
# │ aaa │
# │ bbb │
# │ ccc │
# └─────┘

df3.sort( by=pl.col('key') )
# ┌─────┐
# │ key │
# │ --- │
# │ cat │
# ╞═════╡
# │ aaa │
# │ bbb │
# │ ccc │
# └─────┘

Expected behavior

Column name should not vanish from df2.

Installed versions

---Version info---
Polars: 0.15.11
Index type: UInt32
Platform: macOS-13.1-arm64-arm-64bit
Python: 3.10.9 (main, Dec 15 2022, 17:11:09) [Clang 14.0.0 (clang-1400.0.29.202)]
---Optional dependencies---
pyarrow: 9.0.0
pandas: 1.4.4
numpy: 1.23.3
fsspec: 2022.8.2
connectorx: <not installed>
xlsx2csv: 0.8
matplotlib: <not installed>