SQL - return rows with a duplicate value, where one row in each group has a certain attribute and not another - sql

Question

I am a SQL novice at best - appreciate any help on this answer. It's specific to the program MediaMonkey, but more broadly applicable I think since MediaMonkey's database is I believe SQLite 3. The table I'm querying is called Songs and all the columns I'm referencing are in this table.
I am using the Custom4 field in MediaMonkey to store a song's Work ID from Musicbrainz, and the Custom5 field to store the AcoustID. I'm trying to create a playlist of cover / alternate / live versions of songs that I've rated 5 stars, but where the 5 star song is not a classical work or a jazz recording (I use the Grouping field (GroupDesc) to indicate if a track is classical or jazz).
Parameters for the SQL query need to be something like this:
Same Work ID as a 5 star recording (Rating = 100), where the Grouping
of the 5 star recording is not 'Classical' or 'Jazz' (I am OK with
classical or jazz covers of 5 star popular works)
Different AcoustID from any 5 star recording (eliminate duplicate
recordings)
Rating < 5 stars (I don't want to include the 5 star songs
themselves)
I cobbled together the following SQL query from some examples I found:
Songs.Custom4 IN (SELECT Custom4 FROM Songs GROUP BY Custom4 HAVING Count(*) > 1 AND Max(Rating) > 99 AND Custom4 <> '' AND GroupDesc <> 'Classical' AND GroupDesc <> 'Jazz')
What this appears to include is the following:
Same WorkID as a 5 star recording
Classical or jazz tracks only if the classical or jazz track is rated
5 stars and there is another track with the same Work ID that is not
classical or jazz, or if the classical or jazz track has the same
Work ID as a 5 star recording that is not classical or jazz
Includes duplicates of the 5 star recording, and includes the 5 star
recordings themselves
Somehow then I need to filter out the 5 star classical or jazz tracks and the tracks that have the same Work IDs as those tracks, and I need to exclude the duplicates and the 5 star recordings.
Let's say you have a table like this:
<table class="blueTable">
<thead>
<tr>
<th> Song</th>
<th>Grouping</th>
<th>Rating</th>
<th>Work ID</th>
<th>AcoustID</th>
</tr>
</thead>
<tbody>
<tr>
<td>A1</td>
<td>Classical </td>
<td>100 </td>
<td>a12 </td>
<td>b12 </td>
</tr>
<tr>
<td>A2</td>
<td>Jazz </td>
<td>100 </td>
<td>a21 </td>
<td>b21 </td>
</tr>
<tr>
<td>A3</td>
<td> NULL</td>
<td>100 </td>
<td>a31 </td>
<td>b31 </td>
</tr>
<tr>
<td>A4</td>
<td>NULL </td>
<td>NULL </td>
<td>a12 </td>
<td>b41 </td>
</tr>
<tr>
<td>A5</td>
<td>NULL </td>
<td>NULL </td>
<td>a31 </td>
<td>b31 </td>
</tr>
<tr>
<td>A6</td>
<td>Classical </td>
<td>NULL </td>
<td>a31 </td>
<td>b61 </td>
</tr>
<tr>
<td>A7</td>
<td>NULL </td>
<td>NULL </td>
<td>a31 </td>
<td>b71 </td>
</tr>
</tbody>
</table>
I'd want my query to include the following songs: A6, A7
Here's the logic: A1 and A2 would be excluded since they are Classical and Jazz tracks and are not covers of a popular song with a rating of 100. A3 would be excluded since it has a rating of 100. A4 would be excluded since it's a popular cover of a Classical track that is rated 100. A5 would be excluded since it's a duplicate of A3. A6 is included - it's a Classical cover of A3. A7 is included - it's a cover of A3.