I'm not sure if I've correctly identify the problem: the docx reader might treat the tables as having 1 header row only, while the table cell is empty, the structure is something like this:

```json

[Table [] [AlignDefault,AlignDefault,AlignDefault] [0.0,0.0,0.0]

[[Para [Str "x",Space,Str "y"]]

,[Para [Strong [Emph [Str "a",Space,Str "b"]]]]

,[Para [Strong [Emph [Str "Math"]]]]]

[]]

```

The 2-4th row seems to be a header row, then the `[]` is the table body, which has no length, but should have a length of 3. Panflute asserts that was true, that explains the error I got from my filter.

And pandoc read this just fine, and is indeed output by pandoc's docx reader. On the other hand, pandoc's writers like markdown and html seem to parse this input incorrectly. Is it a valid pandoc AST?

And as a general rule, is it safe to assert that the align-list, width-list, header-list, and each of the row-list are all having the same length?