将行转换为列 Firebird 2.1

Question

我有一个table包含几公里约100 table of kilometers

我需要像这样将行旋转成列 pivot rows into columns

列数固定在行数=10
公里可以重复，有差距，例如：263, 263, 264, 265, 270..

我想用“with recursive”，但是网上的例子很少（如果您能帮我解决问题，我将不胜感激！

Answer 1

正如我在评论中提到的，这可能是在您的表示层中更好地解决的问题，而不是通过查询。我不确定这是否可以通过（递归）CTE 解决，但我可以使用 EXECUTE BLOCK 在 PSQL 中为您提供解决方案（这也可以以存储过程的形式完成）。

此解决方案将行映射到列，并在所有列都已填充后开始新的一行。

execute block
returns (km_1 smallint, km_2 smallint, km_3 smallint, km_4 smallint, km_5 smallint,
  km_6 smallint, km_7 smallint, km_8 smallint, km_9 smallint, km_10 smallint)
as
  declare column_count smallint = 1;
  declare km smallint;
begin
  for select km from kms order by km into km do
  begin
    -- map value to column
    if (column_count = 1) then km_1 = km;
    else if (column_count = 2) then km_2 = km;
    else if (column_count = 3) then km_3 = km;
    else if (column_count = 4) then km_4 = km;
    else if (column_count = 5) then km_5 = km;
    else if (column_count = 6) then km_6 = km;
    else if (column_count = 7) then km_7 = km;
    else if (column_count = 8) then km_8 = km;
    else if (column_count = 9) then km_9 = km;
    else if (column_count = 10) then km_10 = km;
    
    if (column_count < 10) then
      column_count = column_count + 1;
    else
    begin
      -- row complete, output
      column_count = 1;
      suspend;
    end
  end
  if (column_count > 1) then
  begin
    -- partial row
    while (column_count <= 10) do
    begin
      -- null remaining columns
      if (column_count = 2) then km_2 = null;
      else if (column_count = 3) then km_3 = null;
      else if (column_count = 4) then km_4 = null;
      else if (column_count = 5) then km_5 = null;
      else if (column_count = 6) then km_6 = null;
      else if (column_count = 7) then km_7 = null;
      else if (column_count = 8) then km_8 = null;
      else if (column_count = 9) then km_9 = null;
      else if (column_count = 10) then km_10 = null;
      column_count = column_count + 1;
    end
    -- output partial row
    suspend;
  end
end

在 Firebird 4.0 中，您可以使用 window 函数实现相同的结果。

with col_row as (
  select 
    km, 
    mod(row_number() over(order by km) - 1, 10) as column_group, 
    (row_number() over(order by km) - 1) / 10 row_group
  from kms
)
select 
  min(km) filter (where column_group = 0),
  min(km) filter (where column_group = 1),
  min(km) filter (where column_group = 2),
  min(km) filter (where column_group = 3),
  min(km) filter (where column_group = 4),
  min(km) filter (where column_group = 5),
  min(km) filter (where column_group = 7),
  min(km) filter (where column_group = 8),
  min(km) filter (where column_group = 9)
from col_row
group by row_group

对于 Firebird 3.0，将 min(km) filter (where column_group = n) 替换为 min(decode(column_group, n, km)) 或 min(case when column_group = n then km)（其中 n 为 0、1、..、9）。

将行转换为列 Firebird 2.1

Pivot rows into columns Firebird 2.1

sql

pivot

pivot-table

firebird

firebird2.1