Excel - Struggling with complex Index/Match/Match lookup

Question

I've got a table with 9 columns and about 6000 rows. Each row has a price as the last column. Some of those prices are 0.00 when they should be a value.

In another worksheet I have the "original" table with about 3700 rows. The prices I need are in those rows. However, the original table has the prices horizontally within the rows, each next to a cell with specific range of gals. Basically the table I have has unique rows for every location/gal range/price combo, the original has all gal range/prices sequentially in single location rows

For example, a row in the original table looks like this:

... / 1-2000 / 2.8383 / 2001-4000 / 2.5382 / ...

Where as in my new table they look like this:

... / 1-2000 / 2.8383
... / 20001-4000 / 2.5382
etc

Everything is the same in my new table and the original table EXCEPT those gal ranges and prices.

What I'm trying to do is use an array multiple criteria Index/Match (based on 3 cells in both my new table and the original) to lookup the row, find the value that matches the range of gals, and then take the price to the right of that gal range cell.

Here is a row I'm trying to get the value for in my new table:

Here is the row with the value I need in the original table:

Here's a closer look at the formula I've constructed:

INDEX(old!$A$2:$Q$3755,MATCH(1,(A29=old!$A$2:$A$3755)*(F29=old!$F$2:$F$3755)*(G29=old!$G$2:$G3755),1),MATCH(H29,old!$J$2:$Q$3755,1)+1)

The first standard Index/Match part works great... I Index the table and Match to find the row. If I just enter a number for the Col (eg, 1, 2, 3) it'll return the value from the corresponding cell PERFECTLY. However, I can't seem to make the Col match part work... I continuously get REF and N/A errors.

Is there some trick to doing a two way search? It seems like it should be a simple thing to just find that value in the row and take the next cell after if...?

One catch here is that the gal range value I'm looking for is NOT unique... there are at least 20 other references that have the same range (eg, "1-2000"). Is there a way to limit the col match to just the row I find with the row match?

Any help is GREATLY appreciated.

Thanks, Rick

Answer 1

One way to tackle this case is to use #powerquery .

Please refer to this article to find out how to use Power Query on your version of Excel. It is availeble in Excel 2010 Professional Plus and later versions . My demonstration is using Excel 2016 .

The steps are:

1. Load/Add your old data to the Power Query Editor. My sample data only has one row but it is same for thousands of rows;

2. Use Merge Columns function under Transform tab to merge the first 7 columns with a delimiter say semicolon ; ;
3. Repeat Merge Columns for each pair of GALLONS and TOTAL PRICES with the same delimiter ; . If you have done it correctly you should have something like the following:

4. Use Unpivot Columns function under Transform tab to unpivot all the merged columns for GALLONS;TOTAL PRICE , then remove the Attribute column;

5. Use Split Column function under the Transform tab to split each column by delimiter ; . If you have done it correctly you should have something like the following:

6. Make a duplicate column of the GALLONS range column (which is the second last column in the above screen-shot), and then split the original GALLONS range column by delimiter - . Then you should have:

7. Rename the column headers as desired;

8. Close & Load the new table to a new worksheet (by default) or you can change the default setting and create a connection for the new table and load it to a desired location in your workbook.

The second table is the output table and you can do INDEX+MATCH from this new table which should be much easier than from the old table. If the data are identical but just in different structure then you may just use the output table without worrying about looking up missing prices.

I have added a test line to my source table and here is the Refreshed output with a click of button:

Here are the power query M codes behind the scene for reference only. All steps are performed using built-in functions of the editor which is quite straight forward.

let
    Source = Excel.CurrentWorkbook(){[Name="Table1"]}[Content],
    #"Changed Type" = Table.TransformColumnTypes(Source,{{"IATA", Int64.Type}, {"ST", type text}, {"FUELER", type text}, {"UPDATED", type datetime}, {"RESTRICTIONS", type text}, {"BASEF UEL", type text}, {"NOTES", type any}, {"GALLONS1", type text}, {"TOTAL PRICES1", type text}, {"GALLONS2", type text}, {"TOTAL PRICES2", type text}, {"GALLONS3", type text}, {"TOTAL PRICES3", type text}, {"GALLONS4", type text}, {"TOTAL PRICES4", type text}, {"GALLONS5", type text}, {"TOTAL PRICES5", type text}}),
    #"Merged Columns" = Table.CombineColumns(Table.TransformColumnTypes(#"Changed Type", {{"IATA", type text}, {"UPDATED", type text}, {"NOTES", type text}}, "en-AU"),{"IATA", "ST", "FUELER", "UPDATED", "RESTRICTIONS", "BASEF UEL", "NOTES"},Combiner.CombineTextByDelimiter(";", QuoteStyle.None),"Merged"),
    #"Merged Columns1" = Table.CombineColumns(#"Merged Columns",{"GALLONS1", "TOTAL PRICES1"},Combiner.CombineTextByDelimiter(";", QuoteStyle.None),"Merged.1"),
    #"Merged Columns2" = Table.CombineColumns(#"Merged Columns1",{"GALLONS2", "TOTAL PRICES2"},Combiner.CombineTextByDelimiter(";", QuoteStyle.None),"Merged.2"),
    #"Merged Columns3" = Table.CombineColumns(#"Merged Columns2",{"GALLONS3", "TOTAL PRICES3"},Combiner.CombineTextByDelimiter(";", QuoteStyle.None),"Merged.3"),
    #"Merged Columns4" = Table.CombineColumns(#"Merged Columns3",{"GALLONS4", "TOTAL PRICES4"},Combiner.CombineTextByDelimiter(";", QuoteStyle.None),"Merged.4"),
    #"Merged Columns5" = Table.CombineColumns(#"Merged Columns4",{"GALLONS5", "TOTAL PRICES5"},Combiner.CombineTextByDelimiter(";", QuoteStyle.None),"Merged.5"),
    #"Unpivoted Columns" = Table.UnpivotOtherColumns(#"Merged Columns5", {"Merged"}, "Attribute", "Value"),
    #"Removed Columns" = Table.RemoveColumns(#"Unpivoted Columns",{"Attribute"}),
    #"Split Column by Delimiter" = Table.SplitColumn(#"Removed Columns", "Merged", Splitter.SplitTextByDelimiter(";", QuoteStyle.Csv), {"Merged.1", "Merged.2", "Merged.3", "Merged.4", "Merged.5", "Merged.6", "Merged.7"}),
    #"Changed Type1" = Table.TransformColumnTypes(#"Split Column by Delimiter",{{"Merged.1", Int64.Type}, {"Merged.2", type text}, {"Merged.3", type text}, {"Merged.4", type datetime}, {"Merged.5", type text}, {"Merged.6", type text}, {"Merged.7", type text}}),
    #"Split Column by Delimiter1" = Table.SplitColumn(#"Changed Type1", "Value", Splitter.SplitTextByDelimiter(";", QuoteStyle.Csv), {"Value.1", "Value.2"}),
    #"Changed Type2" = Table.TransformColumnTypes(#"Split Column by Delimiter1",{{"Value.1", type text}, {"Value.2", type text}}),
    #"Duplicated Column" = Table.DuplicateColumn(#"Changed Type2", "Value.1", "Value.1 - Copy"),
    #"Reordered Columns" = Table.ReorderColumns(#"Duplicated Column",{"Merged.1", "Merged.2", "Merged.3", "Merged.4", "Merged.5", "Merged.6", "Merged.7", "Value.1 - Copy", "Value.1", "Value.2"}),
    #"Split Column by Delimiter2" = Table.SplitColumn(#"Reordered Columns", "Value.1", Splitter.SplitTextByDelimiter("-", QuoteStyle.Csv), {"Value.1.1", "Value.1.2"}),
    #"Changed Type3" = Table.TransformColumnTypes(#"Split Column by Delimiter2",{{"Value.1.1", Int64.Type}, {"Value.1.2", Int64.Type}}),
    #"Renamed Columns" = Table.RenameColumns(#"Changed Type3",{{"Merged.1", "IATA"}, {"Merged.2", "ST"}, {"Merged.3", "FUELER"}, {"Merged.4", "UPDATED"}, {"Merged.5", "RESTRICTIONS"}, {"Merged.6", "BASEF UEL"}, {"Merged.7", "NOTES"}, {"Value.1 - Copy", "GALLONS"}, {"Value.1.1", "Min Fuel"}, {"Value.1.2", "Max Fuel"}, {"Value.2", "TOTAL PRICE"}}),
    #"Changed Type4" = Table.TransformColumnTypes(#"Renamed Columns",{{"UPDATED", type date}})
in
    #"Changed Type4"

Answer 2

This can be done with Index/Match, but you have to keep your cool.

My Screenshot for reference

The formula in cell F7 is

=INDEX(
INDEX(A2:A3,MATCH(B7&C7&D7,INDEX(A2:A3&C2:C3&F2:F3,0),0))
  :INDEX(L2:L3,MATCH(B7&C7&D7,INDEX(A2:A3&C2:C3&F2:F3,0),0)),
MATCH(E7,
     INDEX(A2:A3,MATCH(B7&C7&D7,INDEX(A2:A3&C2:C3&F2:F3,0),0))
     :INDEX(L2:L3,MATCH(B7&C7&D7,INDEX(A2:A3&C2:C3&F2:F3,0),0)),0)
+1)

To explain: First we build an Index on column A finding the row with a concatenation of the three values. That is combined with the union operator : with an Index on column L, using the same Match.

This Index will return one row of data, from column A to L. It is then used as the range argument for another Index/Match, where inside that row of data Match looks for the "g value" and adds 1 to move one to the right of the found cell.

Note that you don't want whole columns in this formula, since it will be very slow to calculate.

Answer 3

Here is another approach, not using Index/Match at all, but Sumproduct instead. in F7:

=SUMPRODUCT($B$2:$L$3,($A$2:$A$3=B7)*($C$2:$C$3=C7)*($F$2:$F$3=D7)*($A$2:$K$3=E7))

Note that the offset of one column is achieved with the first range from B to L and the last range from A to K.