This is merging DB.SpotifyMeta with DB.raw




Before this point, I should have identified a `wcols`

# ----------------------------- #
Its not clear what I was doing here.  It looks like I was taking a subset of DB.raw
    namely, just dailycounts, by upc, date, artist, mkt_priority
    (just the unqiue rows, so much smaller than DB.raw)
Then merging that *fully* with DB.SpotifyMeta to create DB.merged
# ----------------------------- #


# ----------------------------- #
metcols   <- c("t.DailyCount.byupc", "u.DailyCount.byupc")
colsUsing <- c("upc", "download_date", "artist_country", "mkt_priority", "label_priority", "minDate.byupc", "maxDate.byupc")

setkeyIfNot(DB.SpotifyMeta, "upc")
setkeyIfNot(DB.raw, "upc", "download_date") # confirm this

# DB.merged <- merge(DB.raw[,lapply(.SD, sum, na.rm=TRUE), .SDcols=metcols, keyby=colsUsing] , DB.SpotifyMeta, all=TRUE)

colsUsing <- intersect(colsUsing, names(DB.raw))
dailytotals <- unique( DB.raw[, .SD, .SDcols=c(colsUsing, metcols)] ) 
DB.merged <- merge(dailytotals, DB.SpotifyMeta, all=TRUE)
# ----------------------------- #

## Drop NA startDates
DB.merged <- DB.merged[!is.na(startDate)]

# we dont need this as a stand alone DT, just for easier code reading
Key7Day <- DB.merged[, list(startDate7=seq7Days(startDate)), by=list(upc, startDate)][, list(upc, startDate7)]
simply <- unique(DB.merged[Key7Day, 
                        c(lapply(.SD[, metcols, with=FALSE], sum, na.rm=TRUE), .SD[, !metcols, with=FALSE])
                        , .SDcols=c(metcols, "genre", wcols)])
# ----------------------------- #


as.Date(15720, origin=origin)


DB.merged[CJ(u, )]
is(DB.raw[["download_date"]])

DB.merged[1, c(upc, startDate)]
is(DB.merged[["startDate"]])
DB.merged[1, seq.Date(startDate)]

... then what?