Two series with different filters applied to the same table on the same chart

Jbb
New Member

Hi,

I’d like to draw two series of sells over time :

  • sales of group of customer (a WHERE)
  • a specific customer sales (another WHERE)

So far the best solution I found is to create a view with a derived_table :

view: union_customer_sales_a_b {
derived_table: {
sql:
SELECT
"A" AS selection,
sales_a.date,
sales_a.sell_price
FROM
`customer_sales` AS sales
WHERE
{% condition customer_a %} sales_a.customer_id {% endcondition %}
UNION ALL
SELECT
"B" AS selection,
sales_b.date,
sales_b.sell_price
FROM
`customer_sales` AS sales
WHERE
{% condition customer_b %} sales_b.customer_id {% endcondition %};;
}

parameter: customer_a {
type: unquoted
}

parameter: customer_b {
type: unquoted
}

dimension_group: date {
type: time
timeframes: [
date,
week,
month,
quarter,
year
]
datatype: datetime
sql: ${TABLE}.date ;;
}

dimension: sell_price {
type: number
sql: ${TABLE}.sell_price ;;
}

dimension: selection {
type: string
sql: ${TABLE}.selection ;;
}

measure: total_sell_price {
type: sum
sql: ${sell_price} ;;
}
}

Then in the explore I use the selection as a pivot.

I was wondering if there would be a better pattern to approach this. My concern here is that I’ll have to maintain the base customer_sales view and the composite view with an UNION above.

0 0 81
0 REPLIES 0
Top Labels in this Space