Include zero values in frequency tables #1843
Description
Is your feature request related to a problem? Please describe.
For the "number of results" and "frequency of the search term" visualisations (example), the frequency table will not include rows with a frequency of 0. (In the linked example, these would be years with no matching documents.)
If results are sparse, this results in a lot of missing rows, but these are not missing data; we know that the frequency is 0. If you download the table to create a visualisation in another program, these gaps may not be drawn correctly. (For example, a line graph may show gaps rather than zero values, or worse, draw a straight line from each non-zero value to the next.)
When you make visualisations in Python or R, you can fix this by inserting these values in the data, but this is a hassle. Also, some information is still lost here. In the linked example, the I-analyzer graph starts in 1599, the start of the corpus, but the table starts at the first observation of the search term, in 1660. If I want to make a graph that shows the full search range, I have to manually extend the range of the graph.
You get a similar situation if you compared multiple queries and one of them has no matches; that query will not show up in the CSV. (This may sound like an obscure use case but I did have this problem this week.)
Describe the solution you'd like
Include zero values in the data table.
Describe alternatives you've considered
See above; I had initially written a script to add zero rows to the table after downloading, but there are cases where you still lose information this way.