如何在没有嵌套查询的情况下获取计数并在计算中使用该计数？

Question

我正在尝试获取计数和使用该计数的比率。我尝试过的每个不使用嵌套查询的查询都会给我一个分组错误，因为在计算中使用计数的元组不能按表达式分组。

下面的查询为我提供了我正在寻找的数据。有没有一种方法可以将此查询编写为单个 select 而无需嵌套查询？

select
  org,
  format_number(distinct_preg,0),
  format_number(total_patients,0),
  format_number((distinct_preg/total_patients * 100),2) preg_rate
from (
  select
    preg.org,
    count(distinct preg.patient_id) distinct_preg,
    tot.total_patients total_patients
  from
    pregnancy preg
    join (
      select org, count(distinct patient_id) total_patients from enc group by 1
    ) tot on preg.org = tot.org
  group by 1,3
)
order by 1

---更新----------------------------

这似乎与format_number()函数有关。

此查询有效：

select
  preg.org,
  format_number(count(distinct preg.patient_id),0) distinct_preg,
  tot.total_patients total_patients,
  format_number((count(distinct preg.patient_id) / tot.total_patients * 100),2) preg_rate
from
  pregnancy preg
  join (
    select org, count(distinct patient_id) total_patients from enc group by 1
  ) tot on preg.org = tot.org
group by 1,3

这个给出了显示的错误，仅格式调用不同：

select
  preg.org,
  format_number(count(distinct preg.patient_id),0) distinct_preg,
  format_number(tot.total_patients,0) total_patients,
  format_number((count(distinct preg.patient_id) / tot.total_patients * 100),2) preg_rate
from
  pregnancy preg
  join (
    select org, count(distinct patient_id) total_patients from enc group by 1
  ) tot on preg.org = tot.org
group by 1,3

Answer 1

如果你想要每个患者的平均怀孕次数，你可以使用：

select org, count(*) as num_pregnancies, count(distinct patient_id) as num_patients,
       count(*) * 1.0 / count(distinct patient_id) as avg_pregnancies_per_patient
from pregnancies p
group by org;

没有明确说明您要做什么，这是一种猜测。但是，这似乎是一个非常合理的指标，您正在尝试计算。

Answer 2

显然，编号分组依据是问题所在，如果我将其更改为命名字段，它会起作用。

这个有效：

select
  preg.org,
  format_number(count(distinct preg.patient_id),0) distinct_preg,
  format_number(tot.total_patients,0) total_patients,
  format_number((count(distinct preg.patient_id) / tot.total_patients * 100),2) preg_rate
from
  pregnancy preg
  join (
    select org, count(distinct patient_id) total_patients from enc group by 1
  ) tot on preg.org = tot.org
group by 1,tot.total_patients

这不是：

select
  preg.org,
  format_number(count(distinct preg.patient_id),0) distinct_preg,
  format_number(tot.total_patients,0) total_patients,
  format_number((count(distinct preg.patient_id) / tot.total_patients * 100),2) preg_rate
from
  pregnancy preg
  join (
    select org, count(distinct patient_id) total_patients from enc group by 1
  ) tot on preg.org = tot.org
group by 1,3

如何在没有嵌套查询的情况下获取计数并在计算中使用该计数？

How to get count and use that count in a calculation with out nested query?

sql

apache-spark

databricks