[thread] New Table Feature Announcements

You can now configure dynamic columns from the advanced settings popover. Enabling dynamic columns will create columns for every key in your data source that does not already have a static column defined. Additional format options can also be configured per column format (i.e., number, currency, etc.)

Dynamic Columns

3 Likes

Hello @andoliveyou , where to find the doc for this dynamic columns? thanks.

Excited to have this but, like @AnsonHwang, I don't understand how to use it? I found the option and then there is Label, Format, Hidden, Column Properties (which seems to be same as Format?). Doesn't seem to be anything I'm familiar with from the legacy table.

No Tooltips or docs?

Thanks

@AnsonHwang and @jclutterbuck, thanks for the feedback! We've gone ahead and added some in-app docs which will be available in this week's release.

2 Likes

Any updates here?

Hi everyone! Following up here again to announce some additional table features. We've added:

  1. Aggregation cells can now be rendered in grouped rows! While editing a column, you can select your desired column using the Aggregation option.

  1. You can also enable a summary row in via Table add ons and configure default options for each column by clicking the add on editor. End users will be able to toggle desired aggregations from the summary row component itself.

Screen-Recording-2023-08-08-at-5.16.04-PM

2 Likes

Fantastic! I have two clients that will really like this.

I am playing around with how to implement this for one of them.

When you have more than one group level, it would be super nice to show the count of how many unique items are in the lower level.

I have an imperfect hack. The grouping is on Type and Town text fields. Each Type and Town also has a id and I use unique count on the town id column and call it Town Count. The issue is the Town Count column always shows 1 when it is visible. Makes sense as there is on one of each town. The individual record shows the actual id value. Confusing. Is there a better way?

1 Like

Ah, that's interesting! I don't have a better solution for you at this time.

I can think of a couple paths forwards:

  • We could make the aggregation right of the grouped column's configurable
  • We could introduce a special column setting that enables users to hide non-aggregation data

I'll bring this feedback to the team.

1 Like

A post was split to a new topic: Add more dynamic column settings

3 posts were split to a new topic: Group table by one key & sort by another

We added the json column type to the new table :slightly_smiling_face:

It sound so great. I try it just now, but can't type any character into it, event start with {, what happen?

:thinking: I see what you mean :disappointed: I'll check in with our team about this

Not sure if this has been split off into another topic, but there are some oddities and some feature requests in the aggregation functionality that would be nice to resolve:

  1. Issue: When aggregating by Average, NULL values are interpreted as 0. While possibly valid if a user intends NULL to be interpreted as 0, but usually NULL is meant to be NULL and calculation of average should ignore these records.
    In other words, average(2, 3, 0, 7) = 12 / 4 = 3 while average(2, 3, NULL and 7) = 12 / 3 = 4.
    Perhaps a switch to allow both options, but if not, defaulting to NULL being not counted as an observation instead of NULL = 0 is more appropriate.

  2. It would be useful to add FIRST_VALUE and LAST_VALUE to the default aggregations. This is especially useful for time series (i.e. when grouped, show the final result of a running total column). MIN/MAX doesn't always work here as a field could be declining or increasing over time and therefore the final value might be either the MIN or MAX (or, frankly, none of the above).

  3. I don't have a specific use case for it, but having the ability to aggregate by MODE may also be useful

  4. I'd really like to enable/disable aggregations by level of grouping. For example, it may make sense to sum a field for the lowest level of grouping, but not across the top level of grouping. If I could hide the aggregation at the top level and leave it on for the lower level when expanded, that would be great.

Thanks!

3 Likes

I would also vote for FIRST_VALUE and LAST_VALUE.

I would also like to have a concatenation (CONCAT) summation (tricky I'm sure). But I'm often wanting to summarise text fields, and first/last is okay but sometimes, if it's just a few characters, or just a few entries, it would be nice to have them concatenated?

3 posts were split to a new topic: Item not defined in column status indicator

A post was merged into an existing topic: Item not defined in column status indicator

We've added the 'add row' functionality to the new table component (more info here)

2 Likes

Seconding this:

  1. Issue: When aggregating by Average, NULL values are interpreted as 0. While possibly valid if a user intends NULL to be interpreted as 0, but usually NULL is meant to be NULL and calculation of average should ignore these records. In other words, average(2, 3, 0, 7) = 12 / 4 = 3 while average(2, 3, NULL and 7) = 12 / 3 = 4. Perhaps a switch to allow both options, but if not, defaulting to NULL being not counted as an observation instead of NULL = 0 is more appropriate.

Is there any solution for it?

1 Like

Hey! Would it be possible to for the cell tooltip to show info from another column? For example, if you hover over a cell in the 'User' column, the data in from the relative 'Bio' cell is in the tooltip.