TypeError: Unsupported column type: <class 'numpy.ndarray'>. list or tuple is expected.

[FedericoCeratto]

Hello and thanks for developing the library.
I'm able to reproduce #318 with version 0.2.5 with a minimal example:

CREATE TABLE mytest (                                                           
`status` String                                                                 
)                                                                               
ENGINE = ReplacingMergeTree                                                     
ORDER BY (status)                                                               
SETTINGS index_granularity = 4                                                  

df = pd.DataFrame(columns=["status"])                                           
click.insert_dataframe("INSERT INTO mytest VALUES", df)

A more clear error message would be useful.

[ktncktnc]

I also faced it. No idea what did this message mean.

[xzkostyan]

Client can write and read using:

• usual columns/rows: in lists or tuples
• numpy arrays or pandas dataframes.

In the first case pure client is created without settings={'use_numpy': True} and data for insertion must be provided in lists or tuples.
In the second case numpy client is created with settings={'use_numpy': True} and data for insertion must be provided in numpy arrays of pandas dataframe.

We can make error more clear if we will raise an exception when user calls insert_dataframe/query_dataframe with pure client.