DataFrames can't create tables from UInt8 data

[stensmo]

a=UInt8(1)
 b=UInt8(2)
DataFrame([a;b])

A longer example:

using DataFrames, Parquet

# Download the parquet file
download("https://github.com/vega/vega-datasets/raw/refs/heads/main/data/flights-3m.parquet", "flights-3m.parquet")

# Read the parquet file into a DataFrame
df = DataFrame(Parquet.read("flights-3m.parquet"))

ArgumentError: 'Vector{UInt8}' iterates 'UInt8' values, which doesn't satisfy the Tables.jl AbstractRow interface

[stensmo]

Proposed solution: Add a keyword argument strict, to the DataFrame constructor, which is by default false, where conversions are allowed. Allow anything. strict=true would give the current behavior. Depending on what data format is in the dataset, you need to run df = DataFrame(Parquet.read("flights-3m.parquet")). This may or may not work (depending on the data format). If it does not work, you have to write conversion code. Run it again. Write more conversion code. Then see the data frame. A better workflow would be. See the data in the data frame. Decide that you want to write your own conversion code (enable strict=true).

[stensmo]

This also goes for several other datatypes.

[bkamins]

In your first example you should pass column name:

julia> DataFrame(my_column=[a;b])
2×1 DataFrame
 Row │ my_column
     │ UInt8
─────┼───────────
   1 │         1
   2 │         2

In your second example you use a wrong function from Parquet.jl:

julia> DataFrame(read_parquet("flights-3m.parquet"))
3000000×5 DataFrame
     Row │ date             delay   distance  origin   destination
         │ Int64?           Int64?  Int64?    String?  String?
─────────┼─────────────────────────────────────────────────────────
       1 │ 978307260000000      33      2176  LAS      PHL
       2 │ 978307260000000      19       215  ATL      SAV
       3 │ 978307260000000      14       405  MCI      MDW
       4 │ 978307260000000     -13      2345  ANC      LAX
       5 │ 978307260000000       1        75  RIC      ORF

Thank you for use of DataFrames.jl

[stensmo]

What do you mean? On the first page of Parquet.jl, the first code snippet says this:
using Parquet, DataFrames
df = DataFrame(read_parquet(path))

[bkamins]

You reported using df = DataFrame(Parquet.read("flights-3m.parquet"))