fromFlux should name all of the result columns

[TCL735]

Related to influxdata/ui#3404

According to Flux documentation, annotated CSV, requires all of the header rows and annotation rows. A header row example looks like this: result, table, _start, _stop, _time, _value, _field, _measurement

The result column in the data is always blank. Instead, result relies on the #default annotation row to determine the column name, often defaulted as "_result". In some queries, result may be something else, examples include "mean", "median", "max", "min", "count", and so on. Basically, any name that may come from the yield function in Flux.

Certain visualizations, particularly Band, look for options-specified column names in order to render the graph. Band relies on the user to supply the correct column names, and Giraffe makes the assumption that the user interface would allow the user to correctly identify the column name based on the user's knowledge of the data being analyzed.

However, this is not very user-friendly and requires a certain sophistication, both on the user's part and especially on the implementation details of the user interface.

Proposal:

To assist with identifying the available result columns in annotated CSV, the fromFlux utility should also return all of the result column names, even when there is only a default column name.