MiscTable is not handled correctly by `audformat.Database.update()`

[hagenw]

When combining to databases with a misc table, the misc table should be combined in the end as well.

Minimal example:

import audformat
import pandas as pd


db1 = audformat.Database("mydb")
db1.schemes["answer1"] = audformat.Scheme("str")
db1["sessions"] = audformat.MiscTable(pd.Index(["a"], name="session"))
db1["sessions"]["prompt_1"] = audformat.Column(scheme_id="answer1")
db1["sessions"]["prompt_1"].set(["response1"])

db2 = audformat.Database("mydb")
db2.schemes["answer1"] = audformat.Scheme("str")
db2.schemes["answer2"] = audformat.Scheme("str")
db2["sessions"] = audformat.MiscTable(pd.Index(["b"], name="session"))
db2["sessions"]["prompt_1"] = audformat.Column(scheme_id="answer1")
db2["sessions"]["prompt_2"] = audformat.Column(scheme_id="answer2")
db2["sessions"]["prompt_1"].set(["response1"])
db2["sessions"]["prompt_2"].set(["response2"])

db1.update(db2)

Expected output:

>>> db1["sessions"].df
          prompt_1   prompt_2
session                      
a        response1       <NA>
b        response1  response2

But instead we get:

>>> db1["sessions"].df
          prompt_1   prompt_2
session                      
b        response1  response2

---

This does not happen for a filewise index:

import audformat
import pandas as pd


db1 = audformat.Database("mydb")
db1.schemes["answer1"] = audformat.Scheme("str")
db1["sessions"] = audformat.Table(audformat.filewise_index(["a"]))
db1["sessions"]["prompt_1"] = audformat.Column(scheme_id="answer1")
db1["sessions"]["prompt_1"].set(["response1"])

db2 = audformat.Database("mydb")
db2.schemes["answer1"] = audformat.Scheme("str")
db2.schemes["answer2"] = audformat.Scheme("str")
db2["sessions"] = audformat.Table(audformat.filewise_index(["b"]))
db2["sessions"]["prompt_1"] = audformat.Column(scheme_id="answer1")
db2["sessions"]["prompt_2"] = audformat.Column(scheme_id="answer2")
db2["sessions"]["prompt_1"].set(["response1"])
db2["sessions"]["prompt_2"].set(["response2"])

db1.update(db2)

There we get

>>> db1["sessions"].df
       prompt_1   prompt_2
file                      
a     response1       <NA>
b     response1  response2