Problem using 'crosstalk' to present complex tibble

[FENGQI]

I was trying to present results using rmarkdown document, and to use the crosstalk package to show different part (columns of list of tibble) of the resulting tibble.

Here is a example result tibble

my_example
# A tibble: 9 x 4
Model_Index R.squared Result           summary         
      <dbl>     <dbl> <list>           <list>          
1           1    0.139  <tibble [5 x 5]> <tibble [6 x 7]>
2           2    0.137  <tibble [5 x 5]> <tibble [6 x 7]>
3           3    0.137  <tibble [5 x 5]> <tibble [6 x 7]>
4           4    0.105  <tibble [5 x 5]> <tibble [6 x 7]>
5           5    0.0678 <tibble [5 x 5]> <tibble [6 x 7]>
6           6    0.0677 <tibble [5 x 5]> <tibble [6 x 7]>

I have created the shared data-frame like object by sd <- SharedData$new(my.res) successfully.

Then I'd like to implement is I can pick the model(s) index using filter_slider() , then based on the selection, showing two tables bt DT(), one for Result <tibble [5 x 5]> column, one for summary <tibble [6 x 7]> column.

I have tried put this pipe inside the DT() like

datatable(sd$data() %>% select(Model_Index, Result) %>%
                unnest(cols = c(Result)), extensions="Scroller", 
                style="bootstrap", class="compact",  width="100%",
                options=list(deferRender=TRUE, scrollY=300, scroller=TRUE))
```

Though it did not produce any error, but it gave all models in my results, no matter which models I selected. 

Any help?

Thanks.