Question

我运行两个不同的查询。这个很好用

with t(n) as(
    values
           (1),
           (1),
           (1),
           (2),
           (10)
)
select
       width_bucket(n,array[1,3,6,15]) g,
       count(*)
from t
group by g
order by g

但是当我运行此（真实）查询时，它会在存储桶中报告错误

with data as (
    select
           vendor_id,
           count(distinct pi.id) as cnt
    from payment_invoice_items as pii
    join payment_invoices as pi
        on pii.invoice_id = pi.id
        and pii.deleted_at isnull
    group by vendor_id
)
select
       width_bucket(data.cnt, array[1,10,20,30,40,100,200,400]) as grp,
       count(*)
from data
group by grp
order by grp

data.cnt是一列整数数据。为什么此报告错误？

[42883]错误：函数width_bucket（bigint，integer []）不存在提示：没有函数与给定的名称和参数类型匹配。您可能需要添加显式类型转换。

谢谢

Answer 1

COUNT函数返回类型BIGINT，而不是INTEGER。因此，该类型的错误：

[42883]错误：函数width_bucket（bigint，integer []）不存在提示：没有函数与给定的名称和参数类型匹配。您可能需要添加显式类型转换。

您需要将COUNT的结果显式转换为INTEGER才能在函数中使用：

with data as (
    select
           vendor_id,
           count(distinct pi.id) as cnt
    from payment_invoice_items as pii
    join payment_invoices as pi
        on pii.invoice_id = pi.id
        and pii.deleted_at isnull
    group by vendor_id
)
select
       width_bucket(data.cnt::INTEGER, array[1,10,20,30,40,100,200,400]) as grp,
       count(*)
from data
group by grp
order by grp

Postgres width_bucket：函数width_bucket（bigint，integer []）不存在

1 个答案: