Alternative to a fluent API about arquero HOT 4 CLOSED

I don’t think I fully understand what you are trying to accomplish and why you need params to do it. Can you provide a bit more detail about your task, the input, and the desired output?

from arquero.

Hi @jheer - first thanks for this. It's amazing, I was thinking of building a backend in python to use numpy but this really made prototyping some ideas so much quicker.

I made this observable notebook for you to take a peek, perhaps its clearer? I know how to do the JS bits, was just wondering what how to achieve this in Arquero?

https://observablehq.com/@dscape/alternative-to-a-fluent-api-16

from arquero.

I'm still not entirely sure what transformation logic you have in mind, but here is a guess.

staffAq
  .unroll({ month: d => op.sequence(0, 12) })
  .derive({ monthly: d => op.month(d.start) <= d.month ? d.monthly : 0 })
  .params({ names: ['Jan', 'Feb', 'Mar', 'Apr', 'May', 'Jun', 'Jul', 'Aug', 'Sep', 'Oct', 'Nov', 'Dec'] })
  .derive({ month: (d, $) => $.names[d.month] })
  .groupby('name')
  .pivot('month', 'monthly', { sort: false })
  .select('name', 'Jan', 'Jul', 'Dec')
  .view()

The steps involved are:

1. Add a new array from 0...11 for each month, and unroll it so we get one row per month.
2. Rewrite the monthly column based on the start date.
3. Bind a parameter: an array of month names to use as a lookup table
4. Replace the month indices with month names
5. Group-by the name column to prepare for the next step
6. Pivot the table so that we map values for each month (per-row) to a new columns, one per month
7. (Optionally) Add a select to map the output for your example. If you instead want to filter by some logic, you might include a filter before the pivot, as you see fit.

from arquero.

Thank you, this is excellent!

from arquero.