How to insert from each row into to multiple tables - sql

Question

I'm pretty new to SQL Server (using ssms). I need some help to insert and organize data from one table into multiple tables (which are connected to each other by PK/FK).
The source table has the following columns:
Email, UserName, Phone
It's a messy table with lots of duplicates: same email with different username and so on..
My data tables are:
Person - PersonID(PK, int, not null)
Email - Email (nvarchar, null) , PersonID (FK, int, not null)
Phone - PhoneNumber (int, null) , PersonID (FK, int, not null)
UserName - UserName (nvarchar, null) , PersonID (FK, int, not null)
For each row in the source table, I need to check if the person already exists (by Email); if it does exist, I need to add the new data (if any), else I need to create a new person and add the data.
I searched here for some solutions and find recommendations of using CURSOR.
I tried it, but it takes a really long time to execute (hours.. and still going)
Thanks for any help!
example:
from>
EMAIL | USERNAME | PHONE
------------------------
a#a.a | john | 956165
b#b.b | smith | 123456
c#c.c | bob | 654321
d#d.d | mike | 986514
a#a.a | dan | 658732
e#e.e | dave | 147258
f#f.f | harry | 951962
b#b.b | emmy | 456789
g#g.g | kelly | 789466
h#h.h | kelly | 258369
a#a.a | ana | 852369
to>
EMAIL | PERSONID
----------------
a#a.a | 1
b#b.b | 2
c#c.c | 3
d#d.d | 4
e#e.e | 5
f#f.f | 6
g#g.g | 7
h#h.h | 8
USERNAME | PERSONID
-------------------
john | 1
smith | 2
bob | 3
mike | 4
dan | 1
dave | 5
harry | 6
emmy | 2
kelly | 7
kelly | 8
ana | 1
PHONE | PERSONID
----------------
956165 | 1
123456 | 2
654321 | 3
986514 | 4
658732 | 1
147258 | 5
951962 | 6
456789 | 2
789466 | 7
258369 | 8
852369 | 1

Answer 1

Cursors will generally be slower as they operate on a row-by-row basis. Using set based operations, such as a join, will yield better performance. It's somewhat older, but this article further details the implications of cursors as opposed to set operations. I wasn't entirely sure what columns you want to use to verify matches, as well as what data to add, but a basic example is below and you can fill in the columns as necessary. The Email table was used in the example. For the UPDATE, this will update existing rows based off corresponding rows in the source table. Being an INNER JOIN, only rows with matches on both sides will be impacted. In the second statement, this is an INSERT using only rows from the source table that don't exist in the Email table. This same functionality could also be accomplished using the MERGE statement, however there are a number of issues with this, including problems with deadlocks and key violations.
Update Existing Rows:
UPDATE E
SET E.ColumnA = SRC.ColumnA,
E.ColumnB = SRC.ColumnB
FROM YourDatabase.YourSchema.Email E
INNER JOIN YourDatabase.YourSchema.SourceTable SRC
ON E.Email = SRC.Email
Add New Rows:
INSERT INTO YourDatabase.YourSchema.Email (ColumnA, ColumnB)
SELECT
ColumnA,
ColumnB
FROM YourDatabase.YourSchema.SourceTable
WHERE EMAIL NOT IN ((SELECT EMAIL FROM YourDatabase.YourSchema.Email))