duplicate columns in ascii tables quietly break column name and format parsing

I have found that when a column name is duplicated in an ascii table the column name and format parsing quietly breaks e.g.

This works:

data = """
day precip type
Mon  1.5   rain
Tues 0.0  rain
Wed  1.1 snow
"""

table = ascii.read(data)
table.info()

<Table length=3>
 name   dtype
------ -------
   day    str4
precip float64
  type    str4

If you have a duplicate column name the parsing quietly fails.

data = """
day precip type day
Mon  1.5   rain  Mon
Tues 0.0  rain   Tues
Wed  1.1 snow    Wed
"""

table = ascii.read(data)
table.info()


<Table length=4>
name dtype
---- -----
col1  str4
col2  str6
col3  str4
col4  str3

[pllim]

Since this is an expected feature, can we close the issue?

[taldcroft]

I think that the original issue here, namely reading the file as a different format from expected, has been resolved. io.ascii is doing the correct and expected behavior given the requirement of unique column names.

So I’m closing this, but with the follow-on issue #5374 to consider modifying that requirement and allowing duplicates in the input.