On Fri, 18 May 2012 17:59:12 -0700, Rick Schubert <
r...@san.rr.com> wrote:
>On Fri, 18 May 2012 02:39:21 +1000, RWC wrote:
>
>
>>For any folk mildly interested:
>
>>The Vocal Group Battle artists with the most entries (3 or more):
>
Thanks Rick for reminding me to review 'raw data' before 'posting' - in both
senese of the word :-)
I would estimate, nevertheless, that objectives were well met with, say, 98%
accuracy.
In hindsight, clearly some manual pre-processing adjustments to the raw (pasted
into Notepad from Usenet message) data would have been beneficial - that is, I
should have cast an eye over the entries and, in the light of 'reporting'
objectives, edited some of these contest entries (for example, removed the "-"
from "ABC-Paramount" or inserted a "-" in all "ABC Paramount"s).
>You might want to combine these:
>
>>The Dominoes 9
>>Billy Ward & His Dominoes 4
>
>and
>
>>Hank Ballard & The Midnighters 7
>>The Midnighters 6
>>The Royals 5
>
Gee, this makes a difference in the contest's 'popularity of group ranking'
doesn't it! The only simple (quick) way to fix this situation would be to
manually edit the pre-posted raw data in Notepad, or the after-posted data in
the database table, so as to give a common vocal group name to these 18
recordings.
>
>>Years with the most entries:
>>
>>1955 41
>>1957 41
>
>Roger must not have read this. There were only 40 from those 2 years. You must have counted one
>record twice from each of those years.
>
Yes, I did notice this 41 versus 40 anomaly at the time, but decided to ignore
it :-)
1. for 1955, the doubling up was on song "That's Your Mistake" by
Otis Williams & His Group and
Otis Williams & His New Group
2. for 1957, the doubling up was on artist "The Matadors" with
Vengeance and
Vengeance (Will Be Mine)
This happened because Roger changed the name of
1. vocal group, 2. song title, between the the prelim round and round 1
When combining the 3 Usenet messages that listed contesting pairs of songs
(prelim, round 1 - parts 1 and 2), before posting to the database, I relied on a
specification I made in the database table that artist-song combinations were to
be *unique* - meaning a duplicate artist-song entry would be automatically
rejected (if listed both in prelim and round 1). Roger's changing the name of
song and artist between prelim and round 1 'got around' this filter :-)
Geoff