[CDB-READER] Avoid duplicated `sort_index()`

The following code can and should be optimized by avoiding duplicated `sort_index()`:

```
# set indices for fast lookup
temp_df = (DataFrame(temp_list).set_index(["ID", "LOAD_CASE"], drop=False).sort_index())

if self._data.empty:
   self._data = temp_df
else:
   self._data = concat([self._data, temp_df]).sort_index()
```