BEA fetcher: inconsistency between units in dataset name and units used for series

As reported on the forum https://forum.db.nomics.world/t/problem-in-bea-data-order-of-magnitude/269 there is an inconsistency in units in the data set names. Here

Table 1.1.5. Gross Domestic Product [Billions of dollars] - LastRevised: July 

and the effective units of the time series. The source json https://git.nomics.world/dbnomics-source-data/bea-source-data/blob/master/NIPA/t10105-A.json indicates

    {
      "CL_UNIT":"Level",
      "DataValue":"21,337,898",
      "LineDescription":"Gross domestic product",
      "LineNumber":"1",
      "METRIC_NAME":"Current Dollars",
      "SeriesCode":"A191RC",
      "TableName":"T10105",
      "TimePeriod":"2019Q2",
      "UNIT_MULT":"6"
    },

We is displayed on our web site is therefor 21333898 MILLIONS of current dollars (and not billions as in the dataset name).

The problem comes from the fact that the dataset name indicated in NoteText at the end of the json

  "Notes":[
    {
      "NoteRef":"T10105",
      "NoteText":"Table 1.1.5. Gross Domestic Product [Billions of dollars] - LastRevised: July 26, 2019"
    }

provides inconsistent units.

Proposed solution

1. Report the problem to BEA but it is not sure they want to change it because the note is really about the table name used in their publication.
2. Use fields "METRIC NAME" and UNIT MULT to create a series dimension UNIT
3. Remove the units in the dataset name (in brackets [ ] at the end of the name before the revision) or use LineDescription field
4. Check that it works for all datasets

TODO before closing this issue

• reply to user on forum