寻找最有利可图的范围(数学名称:"maximum subarray problem" 或 "maximum consecutive subsequence sum")
Looking for the most profitable range (mathematical names: "maximum subarray problem" or "maximum consecutive subsequence sum")
为了能够找到最大的 profitable 范围,我将我想要的最低值添加到我想要的最高值,然后我创建了一个 table 就像这个例子:
https://docs.google.com/spreadsheets/d/17zpapBeC5wYxyU6SjbqcbnV4_QP4gooxj0PxdCywDk0/edit?usp=sharing
单元格的公式示例:
在 0
和 0
之间:
=IFERROR(SUM(FILTER($B:$B,($A:$A<=D2)*($A:$A>=$E))))
在 5
和 10
之间:
=IFERROR(SUM(FILTER($B:$B,($A:$A<=D12)*($A:$A>=$J))))
=MAX(E2:O12)
Max Profit = £185.00
=INDEX(A1:O1,ARRAYFORMULA(MIN(IF(E2:O12=MAX(E2:O12),COLUMN(E2:O12)))))
Value Min for Max Profit = 4
=INDEX(D1:D12,ARRAYFORMULA(MAX(IF(E2:O12=MAX(E2:O12),ROW(E2:O12)))))
Value Max for Max Profit = 10
当 A
和 B
中有数百个值¹时,这个 table 变得非常大和沉重,甚至像我当前的原始数据电子表格一样导致崩溃。
有什么方法可以仅使用一个公式或脚本代码来找到 Max Profit | Value Min for Max Profit | Value Max for Max Profit
,而无需逐一计算每个范围,而无需使用数千个单元格,每个单元格都具有特定的公式?
备注:
数百个值¹ → 我的原始电子表格目前在 A
中包含 1471
行数据,结果在 B
中。因此,为了能够进行此分析,我需要在单元格中放入 2,163,841
公式,例如 =IFERROR(SUM(FILTER($B:$B,($A:$A<=D2)*($A:$A>=$E))))
以创建 table 并找到最有利的 table 范围。
最大值:
=INDEX(MAX(IF(SEQUENCE(MAX(A:A)+1)>=SEQUENCE(1, MAX(A:A)+1),
SUMIF(SEQUENCE(MAX(A:A)+1), "<="&SEQUENCE(MAX(A:A)+1), B:B)*
SEQUENCE(1, MAX(A:A)+1, 1, )-QUERY(QUERY(
(SEQUENCE(MAX(A:A)+1)<SEQUENCE(1, MAX(A:A)+1))*B1:B,
"select "&TEXTJOIN(",", 1, "sum(Col"&SEQUENCE(MAX(A:A)+1)&")")),
"offset 1", ), )))
最小值:
=INDEX(REGEXEXTRACT(MAX(IF(SEQUENCE(MAX(A:A)+1)>=SEQUENCE(1, MAX(A:A)+1),
SUMIF(SEQUENCE(MAX(A:A)+1), "<="&SEQUENCE(MAX(A:A)+1), B:B)*
SEQUENCE(1, MAX(A:A)+1, 1, )-QUERY(QUERY(
(SEQUENCE(MAX(A:A)+1)<SEQUENCE(1, MAX(A:A)+1))*B1:B,
"select "&TEXTJOIN(",", 1, "sum(Col"&SEQUENCE(MAX(A:A)+1)&")")),
"offset 1", )+(SEQUENCE(1, MAX(A:A)+1)*10^-10)&9, )*1)&"", "0(\d+)9$")-1)
最大值:
=INDEX(REGEXEXTRACT(MAX(IF(SEQUENCE(MAX(A:A)+1)>=SEQUENCE(1, MAX(A:A)+1),
SUMIF(SEQUENCE(MAX(A:A)+1), "<="&SEQUENCE(MAX(A:A)+1), B:B)*
SEQUENCE(1, MAX(A:A)+1, 1, )-QUERY(QUERY(
(SEQUENCE(MAX(A:A)+1)<SEQUENCE(1, MAX(A:A)+1))*B1:B,
"select "&TEXTJOIN(",", 1, "sum(Col"&SEQUENCE(MAX(A:A)+1)&")")),
"offset 1", )+(SEQUENCE(MAX(A:A)+1)*10^-10)&9, )*1)&"", "0(\d+)9$")-1)
更新:
=INDEX(TEXTJOIN(", ", 1, UNIQUE(FLATTEN(
IF(IF(SEQUENCE(MAX(A:A)+1)>=SEQUENCE(1, MAX(A:A)+1),
SUMIF(SEQUENCE(MAX(A:A)+1), "<="&SEQUENCE(MAX(A:A)+1), B:B)*
SEQUENCE(1, MAX(A:A)+1, 1, )-QUERY(QUERY(
(SEQUENCE(MAX(A:A)+1)<SEQUENCE(1, MAX(A:A)+1))*B1:B,
"select "&TEXTJOIN(",", 1, "sum(Col"&SEQUENCE(MAX(A:A)+1)&")")),
"offset 1", ), )=MAX(IF(SEQUENCE(MAX(A:A)+1)>=SEQUENCE(1, MAX(A:A)+1),
SUMIF(SEQUENCE(MAX(A:A)+1), "<="&SEQUENCE(MAX(A:A)+1), B:B)*
SEQUENCE(1, MAX(A:A)+1, 1, )-QUERY(QUERY(
(SEQUENCE(MAX(A:A)+1)<SEQUENCE(1, MAX(A:A)+1))*B1:B,
"select "&TEXTJOIN(",", 1, "sum(Col"&SEQUENCE(MAX(A:A)+1)&")")),
"offset 1", ), )), SEQUENCE(1, MAX(A:A)+1, 0), )))))
为了能够找到最大的 profitable 范围,我将我想要的最低值添加到我想要的最高值,然后我创建了一个 table 就像这个例子:
https://docs.google.com/spreadsheets/d/17zpapBeC5wYxyU6SjbqcbnV4_QP4gooxj0PxdCywDk0/edit?usp=sharing
单元格的公式示例:
在 0
和 0
之间:
=IFERROR(SUM(FILTER($B:$B,($A:$A<=D2)*($A:$A>=$E))))
在 5
和 10
之间:
=IFERROR(SUM(FILTER($B:$B,($A:$A<=D12)*($A:$A>=$J))))
=MAX(E2:O12)
Max Profit = £185.00
=INDEX(A1:O1,ARRAYFORMULA(MIN(IF(E2:O12=MAX(E2:O12),COLUMN(E2:O12)))))
Value Min for Max Profit = 4
=INDEX(D1:D12,ARRAYFORMULA(MAX(IF(E2:O12=MAX(E2:O12),ROW(E2:O12)))))
Value Max for Max Profit = 10
当 A
和 B
中有数百个值¹时,这个 table 变得非常大和沉重,甚至像我当前的原始数据电子表格一样导致崩溃。
有什么方法可以仅使用一个公式或脚本代码来找到 Max Profit | Value Min for Max Profit | Value Max for Max Profit
,而无需逐一计算每个范围,而无需使用数千个单元格,每个单元格都具有特定的公式?
备注:
数百个值¹ → 我的原始电子表格目前在 A
中包含 1471
行数据,结果在 B
中。因此,为了能够进行此分析,我需要在单元格中放入 2,163,841
公式,例如 =IFERROR(SUM(FILTER($B:$B,($A:$A<=D2)*($A:$A>=$E))))
以创建 table 并找到最有利的 table 范围。
最大值:
=INDEX(MAX(IF(SEQUENCE(MAX(A:A)+1)>=SEQUENCE(1, MAX(A:A)+1),
SUMIF(SEQUENCE(MAX(A:A)+1), "<="&SEQUENCE(MAX(A:A)+1), B:B)*
SEQUENCE(1, MAX(A:A)+1, 1, )-QUERY(QUERY(
(SEQUENCE(MAX(A:A)+1)<SEQUENCE(1, MAX(A:A)+1))*B1:B,
"select "&TEXTJOIN(",", 1, "sum(Col"&SEQUENCE(MAX(A:A)+1)&")")),
"offset 1", ), )))
最小值:
=INDEX(REGEXEXTRACT(MAX(IF(SEQUENCE(MAX(A:A)+1)>=SEQUENCE(1, MAX(A:A)+1),
SUMIF(SEQUENCE(MAX(A:A)+1), "<="&SEQUENCE(MAX(A:A)+1), B:B)*
SEQUENCE(1, MAX(A:A)+1, 1, )-QUERY(QUERY(
(SEQUENCE(MAX(A:A)+1)<SEQUENCE(1, MAX(A:A)+1))*B1:B,
"select "&TEXTJOIN(",", 1, "sum(Col"&SEQUENCE(MAX(A:A)+1)&")")),
"offset 1", )+(SEQUENCE(1, MAX(A:A)+1)*10^-10)&9, )*1)&"", "0(\d+)9$")-1)
最大值:
=INDEX(REGEXEXTRACT(MAX(IF(SEQUENCE(MAX(A:A)+1)>=SEQUENCE(1, MAX(A:A)+1),
SUMIF(SEQUENCE(MAX(A:A)+1), "<="&SEQUENCE(MAX(A:A)+1), B:B)*
SEQUENCE(1, MAX(A:A)+1, 1, )-QUERY(QUERY(
(SEQUENCE(MAX(A:A)+1)<SEQUENCE(1, MAX(A:A)+1))*B1:B,
"select "&TEXTJOIN(",", 1, "sum(Col"&SEQUENCE(MAX(A:A)+1)&")")),
"offset 1", )+(SEQUENCE(MAX(A:A)+1)*10^-10)&9, )*1)&"", "0(\d+)9$")-1)
更新:
=INDEX(TEXTJOIN(", ", 1, UNIQUE(FLATTEN(
IF(IF(SEQUENCE(MAX(A:A)+1)>=SEQUENCE(1, MAX(A:A)+1),
SUMIF(SEQUENCE(MAX(A:A)+1), "<="&SEQUENCE(MAX(A:A)+1), B:B)*
SEQUENCE(1, MAX(A:A)+1, 1, )-QUERY(QUERY(
(SEQUENCE(MAX(A:A)+1)<SEQUENCE(1, MAX(A:A)+1))*B1:B,
"select "&TEXTJOIN(",", 1, "sum(Col"&SEQUENCE(MAX(A:A)+1)&")")),
"offset 1", ), )=MAX(IF(SEQUENCE(MAX(A:A)+1)>=SEQUENCE(1, MAX(A:A)+1),
SUMIF(SEQUENCE(MAX(A:A)+1), "<="&SEQUENCE(MAX(A:A)+1), B:B)*
SEQUENCE(1, MAX(A:A)+1, 1, )-QUERY(QUERY(
(SEQUENCE(MAX(A:A)+1)<SEQUENCE(1, MAX(A:A)+1))*B1:B,
"select "&TEXTJOIN(",", 1, "sum(Col"&SEQUENCE(MAX(A:A)+1)&")")),
"offset 1", ), )), SEQUENCE(1, MAX(A:A)+1, 0), )))))