GROUP rows when conditions are met - mysql

Question

I hope you can help me out with the next problem. I am trying to figure out how to count many records exist when some conditions occur.
An example of my input is shown below:
fullname emailaddress1 telephone1
Juana Foster juana
Juana Foster juana 1933595322
Henzley 1841901633
Henzley henzley 1841901633
Hoyle hoyle 1584190699
Hoyle hoyle 1584190655
Aaron Jeans alpha2222 1816808600
Aaron Jeans alpha 1816808600
Erick Martin a1009 1816250211
Erick Martin martin 1565960141
Erick Martin a1009
Erick Martin martin 1565960141
I would like to group the occurrences which match the next conditions:
Fullname = Fullname and ((emailaddress = emailaddress and emailaddress != '') OR (telephone=telephone and telephone != '')).
In other words, I want to group in one row, all the rows which have the same fullname and, same email or address. Email and Address must be distinct to empty to be considered matchable.
The expected output would be:
fullname occurrences
Juana Foster 1
Henzley 1
Hoyle 1
Aaron Jeans 1
Erick Martin 2
I don't have any problem to solve the problem in a loop, but I have been thinking how to solve the problem in SQL, and tried GROUP BY and UNIONS, however and I haven't reached the solution. I am using mySQL.
UPDATE:
I provide a new example with more specific cases, in order to clarify the information:
For example for the next input
fullname emailaddress1 telephone1
Aaron Jeans alpha2222 1816808600
Aaron Jeans 1816808600
Aaron Jeans alpha2222 1816808600
Aaron Jeans alpha 1816808600
Erick Martin a1009 1816250211
Erick Martin a1009
Erick Martin 1816250211
Erick Martin martin 1565960141
Erick Martin martin 1565960141
Nacho Mason 1111111111
Nacho Mason 2222222222
In this case the output should be:
Aaron Jeans 1
Erick Martin 2
Nacho Mason 2
Aaron Jeans has 1 occurrence because his 4 records share the same telephone.
Erick Martin has 2 occurrences, the first one is for the next cases:
Erick Martin a1009 1816250211
Erick Martin a1009
Erick Martin 1816250211
Because the 3 records share same fullname and (same email(a1009) or same phone [phone]), these three records are considered 1 occurrence.
The second occurrence for Erick Martin match with the next two records, because it has same fullname, same email and same phone.
Erick Martin martin 1565960141
Erick Martin martin 1565960141
Nacho Mason has 2 occurrences, because he has 2 different phones, and as his email is empty, and can't be considered equal.

Answer 1

I think this might be what you're trying to do.
first you want to do a SELECT of all occurrences where Email is not blank and telephone is not blank like
SELECT fullname,emailaddress1,telephone1,COUNT(*) AS occurrences
FROM T
WHERE emailaddress1 != ''
AND telephone1 != ''
GROUP BY fullname,emailaddress1,telephone1
then now you outer SELECT and MAX(occurrences) GROUP BY fullname
SELECT fullname, MAX(occurrences) as occurrences
FROM
(SELECT fullname,emailaddress1,telephone1,COUNT(*) AS occurrences
FROM T
WHERE emailaddress1 != ''
AND telephone1 != ''
GROUP BY fullname,emailaddress1,telephone1
)AS result
GROUP BY fullname
sqlfiddle
At least, I think that's what you're trying to get at.
curious though, what if you had another full name with 3 rows of identical email and phone number and the same full name with another 2 rows of identical email and phone number like this
Adam Smith adam 1234
Adam Smith adam 1234
Adam Smith adam 1234
Adam Smith smith 5678
Adam Smith smith 5678
Do you want to show as Adam Smith 3 or Adam Smith 2 or both?
the query above will you Adam Smith 3
UPDATE: I guess based on your desired output you want Adam Smith 3 because with your Erick Martin 2, he would've had Erick Martin 1, Erick Martin 2 from the inner select.