Joel Buhr
asked on
Count Distinct Records by Address MySQL
I have a database where I know that I may have more than one record present for a given Address, Individual, or Surname+Address (HOUSEHOLD)
The data in the table is standardized. Fields include :
EMAIL
FIRST NAME
LAST NAME
ADDRESS
CITY
STATE
ZIPCODE
URL
CODE
I need to be able to run a query by either URL, CODE, State, AND ZIP
For Example
or
But I need to return a count or even pull the data based on ONE record per Household (household being defined as : Surname, Address, City, State, ZIP), or one per Address (Address being defined as : Address, City, State, ZIP)
How would I do this in combination with the other selects
The data in the table is standardized. Fields include :
FIRST NAME
LAST NAME
ADDRESS
CITY
STATE
ZIPCODE
URL
CODE
I need to be able to run a query by either URL, CODE, State, AND ZIP
For Example
WHERE URL LIKE '%motorcycle%'
and STATE in ('mi', 'oh')
GROUP BY STATE
or
WHERE CODE IN ('abcd', 'xyz')
GROUP BY STATE
But I need to return a count or even pull the data based on ONE record per Household (household being defined as : Surname, Address, City, State, ZIP), or one per Address (Address being defined as : Address, City, State, ZIP)
How would I do this in combination with the other selects
ASKER CERTIFIED SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
We have had to deal with a very similar issue in our constituent management app. The problem boils down to two things:
Hope this helps!
There are an infinite number of ways to write a single address. Between typos, non-standard abbreviations, and just plain wrongness, you can never be sure that any two addresses point to the same place. We work around this (to a degree) with an in-house application which leverages USPS official address lists and standardization techniques. Every address we receive is initially sent through this app for formatting and term standardization, and verification of physical existence (according to the USPS, anyways).
Your firm definition of "household" is a good thing, but a bit complicated to be using "on the fly", so to speak. It might be better to abstract this information to a "household identifier" table. For example, you could move any address-specific fields in YourTable to AddressIDTable, replacing them with the ID field from the new lookup. This would allow for a much simpler comparison to decide if two people are living at the same address - if the IDs match, then yes.
Hope this helps!
ASKER
Zberteoc
Thank you for your response I will try that. The examples is what I was looking for.
Steve Bink
Very helpful suggestions. I will look into this.
James0628
Although I appreciate your response to this question. I do realize that if I want one per household that it is a combination of a few fields. What I did not know how to do is create the correct query. To answer your question. We are searching by URL or Code to select specific people in the table by what code has been assigned to them or by the URL they came in on. Its one thing to simply write the query for that, but another to limit the results to one per Address or one per Household in combination with the other.
Thank you for your response I will try that. The examples is what I was looking for.
Steve Bink
Very helpful suggestions. I will look into this.
James0628
Although I appreciate your response to this question. I do realize that if I want one per household that it is a combination of a few fields. What I did not know how to do is create the correct query. To answer your question. We are searching by URL or Code to select specific people in the table by what code has been assigned to them or by the URL they came in on. Its one thing to simply write the query for that, but another to limit the results to one per Address or one per Household in combination with the other.
Since you were already using Group By, adding more fields to that seemed straightforward enough (which made me think that there might be more to the question).
As for the counts, I was originally thinking of counting the records within each Household/Address (eg. to identify the Households that have multiple records). If you want to count the number of unique Households/Addresses, you can combine the fields that make up a Household/Address and do a count on that. For example, for a Household count, assuming that ZIP is a numeric field, it might be something like
COUNT (DISTINCT Surname, Address, City, State, CAST (ZIP AS CHAR)) AS Count_Households
James
As for the counts, I was originally thinking of counting the records within each Household/Address (eg. to identify the Households that have multiple records). If you want to count the number of unique Households/Addresses, you can combine the fields that make up a Household/Address and do a count on that. For example, for a Household count, assuming that ZIP is a numeric field, it might be something like
COUNT (DISTINCT Surname, Address, City, State, CAST (ZIP AS CHAR)) AS Count_Households
James
I'm not sure how the two different Where's (checking URL and STATE, or checking CODE), or the two different "groups" (Household or Address) fit in. Is the user going to be selecting which fields they're testing and which "group" they're looking for? Are you trying to handle all of those possibilities in one query?
James