Implementing the recursion using flux Query

InfluxDB 2
,
1

As I have my data in different buckets based on the month and year.
So my data will be in the different buckets as follows:
For Jan 2023 in 012023
Dec 2022 in 122022
Nov 2022 in 112022
so if I query the data for large duration then data will be present in multiple buckets.
After getting the month and year from the start and stop time range I need to generate the bucket name dynamically and aggregate them and give the result.
The first approach will be using loops which going to be difficult to do in flux.
So I thought we can do this using Recursion.

I have written the first code below to achieve the recursion


import "date"
start_month = int(v: date.month(t: v.timeRangeStart))
end_month = int(v:date.month(t: v.timeRangeStop ))

diff = end_month - start_month
diff_duration = if diff < 0 then end_month - start_month + 12 else end_month - start_month
start_year = int(v: date.year(t: v.timeRangeStart))
end_year = int(v:date.year(t: v.timeRangeStop ))

bucket_name =  string(v: start_year) + string(v:start_month)
f = (input_bucket) =>
	( if diff_duration != 0 then f(input_bucket: bucket_name).unused else die(msg: "script stopped")
  
  from(bucket: input_bucket)
  |> range(start: v.timeRangeStart, stop: v.timeRangeStop)
  |> filter(fn: (r) => r["_measurement"] == "measurement")
  |> filter(fn: (r) => r["tag1"] == "tag_value")
  |> filter(fn: (r) => r["_field"] == "field_name")
  |> aggregateWindow(every: 1d, fn: last, createEmpty: false)
  |> map(fn: (r) => ({ 
//       r with
      _value:r._value,
      _field:r._field,
      start_duration:v.timeRangeStart,
      end_duration:v.timeRangeStop,
      "bucket_name": bucket_name,
      "start_month":start_month,
      "end_month":end_month, 
      "diff_duration": diff_duration,
      "diff": diff,
     }))
  |> yield(name: "last")
  )

 f( 	input_bucket: bucket_name)

But this code giving the error as unidentified unidentified f at line if diff_duration != 0 then f(input_bucket: bucket_name).unused else die(msg: "script stopped")

I tried a different approach to achieve recursion using a different approach

import "date"
start_month = int(v: date.month(t: v.timeRangeStart))
end_month = int(v:date.month(t: v.timeRangeStop ))

diff = end_month - start_month
diff_duration = if diff < 0 then end_month - start_month + 12 else end_month - start_month
start_year = int(v: date.year(t: v.timeRangeStart))
end_year = int(v:date.year(t: v.timeRangeStop ))

bucket_name =  string(v: start_year) + string(v:start_month)

// g = (input_bucket) =>(
// 

g = (input_bucket) =>{
unused = from(bucket: input_bucket)
  |> range(start: v.timeRangeStart, stop: v.timeRangeStop)
  |> filter(fn: (r) => r["_measurement"] == "measurement")
  |> filter(fn: (r) => r["tag1"] == "tag_value")
  |> filter(fn: (r) => r["_field"] == "field_name")
  |> aggregateWindow(every: 1d, fn: last, createEmpty: false)
  |> map(fn: (r) => ({ 
//       r with
      _value:r._value,
      _field:r._field,
      start_duration:v.timeRangeStart,
      end_duration:v.timeRangeStop,
      "bucket_name1": bucket_name,
      "start_month1":start_month,
      "end_month1":end_month, 
      "diff_duration1": diff_duration,
      "diff1": diff,
     }))
  return {  unused: unused }

  |> yield(name: "last")
  }





f = (input_bucket) =>
	( if diff_duration != 0 then g(input_bucket: bucket_name).unused else die(msg: "script stopped")
  
  from(bucket: input_bucket)
  |> range(start: v.timeRangeStart, stop: v.timeRangeStop)
  |> filter(fn: (r) => r["_measurement"] == "measurement")
  |> filter(fn: (r) => r["tag1"] == "tag_value")
  |> filter(fn: (r) => r["_field"] == "field_name")
  |> aggregateWindow(every: 1d, fn: last, createEmpty: false)
  |> map(fn: (r) => ({ 
//       r with
      _value:r._value,
      _field:r._field,
      start_duration:v.timeRangeStart,
      end_duration:v.timeRangeStop,
      "bucket_name": bucket_name,
      "start_month":start_month,
      "end_month":end_month, 
      "diff_duration": diff_duration,
      "diff": diff,
     }))
  |> yield(name: "last")
  )

 f(
 	input_bucket: bucket_name
)


//g(
//	input_bucket: bucket_name
//).unused

I am getting this error

type error @36:6-36:25: expected [A] but found {unused:[{start_month1:int, start_duration:time, end_month1:int, end_duration:time, diff_duration1:int, diff1:int, bucket_name1:string, _value:C, _field:B}]} (argument tables)

@Anaisdg @Jay_Clifford
Is there any approach where I can implement the same?

2

@Anaisdg @scott could please help me with this?
Thanks

3

@Ravikant_Gautam Is there a reason you’re storing data like this (bucket per month)? This certainly complicates querying the data. Generally buckets are differentiated by their data retention and data precision.

4

When we keep sending our data into one bucket then data is increasing which reduces the query performance and increase then querying time.
So we got a approach to have data into multiple buckets based on month and year.
When someone query the data we fetch the data based start and stop time range.

5

I’ve never seen this behavior. If this is true, it’s a bug an should be reported. The amount of data in a bucket shouldn’t affect how queries perform against that bucket unless you’re querying everything in the bucket. But most queries don’t need to do that.