Categorize rows based on value in other columns

Hi Plurp , you can consider using Text.Contains search and Text.AfterDelimiter exact the Category, and the fill down to achieve the result.  Then remove the rows containing Category and Date.

let
    Source = Table.FromRows(Json.Document(Binary.Decompress(Binary.FromText("i45Wck4sSU3PL6pUMDK3UgjJryxW0lFSitWJVnIBSsDYhob6hvpGBkbGQAEVcz0DA5ioEVzUEEnUGC5qDBMNyS9JzFFIyy9SQLIQrA+uES5hamCl4Jafn4LNJRZIhpvBdFromyDcYYQQNYWLWiIEzTA9gsVxpgZg55uAlcQCAA==", BinaryEncoding.Base64), Compression.Deflate)), let _t = ((type nullable text) meta [Serialized.Text = false]), _n = (( Currency.Type ) meta [Serialized.Text = false]) in type table [Column1 = _t, Cost = _n]),
    #"Added Custom" = Table.AddColumn(Source, "Category", each if Text.Contains( [Column1] , ": ") then Text.AfterDelimiter( [Column1] , ": ") else null, type text),
    #"Filled Down" = Table.FillDown(#"Added Custom",{"Category"}),
    #"Filtered Rows" = Table.SelectRows(#"Filled Down", each not Text.Contains([Column1], "Category") and not Text.Contains([Column1], "Date")),
    #"Renamed Columns" = Table.RenameColumns(#"Filtered Rows",{{"Column1", "Date"}}),
    #"Parsed Date" = Table.TransformColumns(#"Renamed Columns",{{"Date", each Date.From(DateTimeZone.From(_)), type date}}),
    #"Extracted Text After Delimiter" = Table.TransformColumns(#"Parsed Date", {{"Cost", each Text.AfterDelimiter(Text.From(_, "en-GB"), "$"), type text}}),
    #"Changed Type" = Table.TransformColumnTypes(#"Extracted Text After Delimiter",{{"Cost", Currency.Type}})
in
    #"Changed Type"