[FEA] Implicit typecasting between Decimal and Floats during concat

[ayushdg]

Is your feature request related to a problem? Please describe.
Concat automatically handles typecasting between to decimal columns to decide the precision and scale of the output resulting decimal column during concat. There is implicit typecasting between other numeric column types as well (float32/64 and ints). However it isn't possible to concat between float and decimal columns in a dataframe.

Describe the solution you'd like

df1['a'] = range(10)
df1['b'] = 1.234
df1.b = df1['b'].astype(Decimal64Dtype(7,2))

df2['b'] = range(10)
df2['a'] = 1.234
cudf.concat([df1, df2])

ValueError: All columns must be the same type

df1['a'] = range(10)
df1['b'] = 1.234
df1.b = df1['b'].astype("float32")

df2['b'] = range(10)
df2['a'] = 1.234
cudf.concat([df1, df2])
# Works

Describe alternatives you've considered
Alternative is to cast explicitly before concat

Additional context
N/A