Power Query: Merging on virtual table from starting CSV takes too long and repeats code execution

I have a starting CSV that contains a SELECT DISTINCT of all the dimension columns of my transaction data. The columns are hierarchical, so I am trying to split that further into three tables based on column distribution and frequency of reporting level required.

My code is similar to the following, where I create an index on a distinct combination of the first set of columns and merge it back to the original master. After repeating twice for two other sets of columns, I merge the original master id to the transaction table master id, to yield the three dimension foreign keys for a more defined star schema.

let
    Source = Folder.Files(#"Folder Path"),
    ReadCSV = Table.AddColumn(Source, "CSV", each
                    Table.PromoteHeaders(
                                            Csv.Document([Content], [Delimiter=",", Encoding=1252, QuoteStyle=QuoteStyle.None]), [PromoteAllScalars=true]
                                        )
                ),
    KeepCSV = Table.SelectColumns(ReadCSV,{"CSV"}),
    ExpandedCSV = Table.ExpandTableColumn(
                            KeepCSV, "CSV",
                            {"list", "of", "columns"},
                            {"List", "Of", "columns"}
                        ),
    ChangedType = Table.TransformColumnTypes(ExpandedCSV,{{"List", type text}, {"Of", type text}, {"Columns", type text}}),
	
	/* Lines 17 - 24 (EventScenarios - ExpandedES) is repeated similarly two more times, with some other columns */
    EventScenarios = let
                        ESSource = Table.Distinct( Table.Group(ChangedType, {"GroupColumn"}, {{"ValueColumn", each List.Max([ValueColumn])}}) ),
                        ESSorted = Table.Sort(ESSource, {{"GroupColumn", Order.Ascending}}),
                        ESFinal = Table.AddIndexColumn(ESSorted, "Id", 1, 1, Int64.Type)
                     in
                        ESFinal,
    MergedES = Table.NestedJoin(ChangedType, {"Event", "Group", "Scenario"}, EventScenarios, {"Event", "Group", "Scenario"}, "Event Scenarios", JoinKind.LeftOuter),
    ExpandedES = Table.ExpandTableColumn(MergedES, "Event Scenarios", {"Id"}, {"Event Scenario Id"})
in
    ExpandedES

The code above works. However, it doesn't seem to cache the previous merge operations nor the inner generation of distinct rows.

The original file is only 65mb (uncompressed), but the loaded data increases to around 570mb.

Since the dimensions are hierarchical and I include the previous generated master table's ID, in the subsequent select distinct, Power Query seems to re-process the previously generated tables again. I tried enclosing them in Table.Buffer, but that only seemed to make things worse.

Can someone suggest how may I improve the loading process.