Check Duplicated records in a table - sql

I have a table storing user's information, including SSN. Now I need to write a query to prevent, or to check, when there is person who start to insert some new record into the table, if the new records' SSN is the same with someone who currently in the system, then it return false, or reject the insert.
I know how to count same SSN in a table, but this is sort of before insert check, should I do this on the front end level?
Update: So I think through this again, I am asking about, how can I actually allow the insert, and also every time we do the insert, there will be a check, if the new insert have duplicate SSN with one of our current client, then there will be a new column called "DuplicatedSSN"=True. How can I achieve this?

Step 1 (DB): Add a UNIQUE constraint to the table to prevent duplicates.
Step 2 (program): Do a pre-check against the value and run your code from there.
Pseudocode would look like:
SELECT COUNT(*) LINES FROM MAIN_TABLE WHERE SSN = $attempted_entry
if(LINES <= 0){
--insert into MAIN_TABLE
} else {
--insert into POSSIBLE_FRAUD_CHECKS
}

I hope you are encrypting SSN information, but here's how you check if there's a duplicate already.. There's a couple ways.
First you should make the database column for SSN have a 'UNIQUE' value. This will prevent there ever being a duplicate.
With your query you should do the following:
INSERT IGNORE INTO table (column1, column2, etc, ssn_column)
VALUES (column1_value, column2_value, etc_value, ssn_value)
ON DUPLICATE KEY UPDATE ssn_column=ssn_column
Basically this says...
1) INSERT values into table
2) BUT if the ssn already exists...
3) Update the existing ssn to = ssn (aka do nothing)
This will only insert a new column if it doesn't exist already, but SSN column MUST be a UNIQUE column!
You can also do...
SELECT id FROM table WHERE ssn = ssn_value
if this returns a row... you know it exists...
if nothing exists...
INSERT INTO table (columns) VALUES (values)

Alex is correct here. A UNIQUE constraint will prevent duplicates entries in your database. In order to add the constraint, you will need to make sure there aren't any duplicates existing in your table. To add the constraint you could use something like the following
ALTER TABLE table1 ADD UNIQUE (SSN)
If you want to just set a field based on existing information, I think you can use a case statement in your insert.
INSERT INTO table 1 (SSN, DuplicatedSSN) VALUES (999-99-9999, CASE WHEN EXISTS (SELECT ID FROM orders where SSN = '999-99-9999') THEN 1 ELSE 0 END)

Related

Insert row to database based on form values not currently in database

I am using Access 2013 and I am trying to insert rows to a table but I don't want any duplicates. Basically if not exists in table enter the data to table. I have tried to using 'Not Exists' and 'Not in' and currently it still does not insert to table. Here is my code if I remove the where condition then it inserts to table but If I enter same record it duplicates. Here is my code:
INSERT INTO [UB-04s] ( consumer_id, prov_id, total_charges, [non-covered_chrgs], patient_name )
VALUES ([Forms]![frmHospitalEOR]![client_ID], [Forms]![frmHospitalEOR]![ID], Forms![frmHospitalEOR].[frmItemizedStmtTotals].Form.[TOTAL BILLED], Forms![frmHospitalEOR].[frmItemizedStmtTotals].Form.[TOTAL BILLED], [Forms]![frmHospitalEOR]![patient_name])
WHERE [Forms]![frmHospitalEOR]![ID]
NOT IN (SELECT DISTINCT prov_id FROM [UB-04s]);
You cannot use WHERE in this kind of SQL:
INSERT INTO tablename (fieldname) VALUES ('value');
You can add a constraint to the database, like a unique index, then the insert will fail with an error message. It is possible to have multiple NULL values for several rows, the unique index makes sure that rows with values are unique.
To avoid these kind of error messages you can build a procedure or use code to check data first, and then perform some action - like do the insert or cancel.
This select could be used to check data:
SELECT COUNT(*) FROM [UB-04s] WHERE prov_id = [Forms]![frmHospitalEOR]![ID]
It will return number of rows with the spesific value, if it is 0 then you are redy to run the insert.

Relating two tables

I have created tables T1 with columns( id as Primary key and name) and T2 with columns( id as primary key, name, t_id as foreign key references T1(id)) . I Inserted some values from inputs from a Windows form. After querying SELECT * FROM T2; using isql, all the values in the foreign key column are null instead of duplicating values in T1(id) because of the relationship created. Is they anything I have left out or need to add? The primary key of both tables are autoincremented.
You are confusing auto-incremented keys and relationship uses.
Auto-incremented keys (or generally talking, fields) just help you when you are inserting a new record on the table of the key. But when you are inserting a new record that makes a reference to a record in another table, then you must specify that record, using the foreign key field. Or in your case, the user that is inserting the "name" in T2 must say which one record on T1 that "name" in T2 is making a reference.
Your confusion on the relationship is that you are thinking that an established relationship will enforce the use of that values automatically. But the relationship just enforce the validation of the values. So, the field t_id in T2 will not use the value of the last record of T1 automatically. But if you try to insert a value that do not exist in T1 in the field t_id, the relationship will not let you do.
So, answering your question, what you left out and need to add?
You left out the part of the code that insert the value on the t_id field of T2 table.
Let me try to explain using an example that is more common.
The most common case of this is that the application insert first the T1 record and then when the user is inserting T2, the application provide a way to the user to choose which one T1 record his T2 record is referencing.
Suppose T1 is a publishers table and T2 is a book table. User insert a publisher, and when it is inserting a book it can choose which one publisher publish that book.
Field "ID" of Customers will be AUTOINCREMENT by default in table create using Event BeforeInsert on table CUSTOMERS. LOOK AT
CREATE TRIGGER nametrigger FOR nametable
ACTIVE BEFORE INSERT POSITION 0
AS
BEGIN
IF (NEW.ID IS NULL) THEN BEGIN
NEW.ID = GEN_ID(GEN_PK_ID, 1);
END
END
Now one new record in Customers
INSERT INTO Customers (CustomerName, ContactName, Address, City, PostalCode, Country)
VALUES ('Cardinal','Tom B. Erichsen','Skagen 21','Stavanger','4006','Norway');
Then ID will be automaticaly one sequencial number from 1 up to last integer or smallint or bigint as you defined in your create table (pay attencion that ID field is not include in FIELDS and VALUES) because TRIGGER
now you can use the dataset (obj) options to link the table MATER and DETAIL see in help delphi
or in SQL you can to use PARAMS FIELDS
later insert one new record in table MASTER try...
INSERT INTO xTable2 (IDcustomersField, ..., ..., ...., ....)
VALUES ( :IDcustomersField, ..., ..., ...., ....);
xTable2 may using one field ID (Primary Key) autoincrement too. this help when DELETING or UPDATING fileds in this table
Then you can say the value to :IDcustomersField in table detail using
xQuery.PARAM( 0 ).value or xQuery.PARAMBYNAME( IDcustomersField).value (here im using Query obj as example )
you can to use example with DATASOURCE in code to say the value for IDcustomersField
can to use
Events in SQL
can to use
PROCEDURE IN SQL
DONT FORGOT
you have to create Relationship between two table ( REFERENCIAL INTEGRITY and PRIMARY KEY in mater table ) NOT NULL FOR TWO FIELDS ON TABLES
I believe that understand me about my poor explanation (i dont speak english
You need to insert the values for t_id manually, after you get the ID's value from the main table T1.
Depending on your logic in the database you also can use a trigger or a stored procedure. Give us more information about what values you expect to have in NAME field in T2 after the insert? Are they duplicates from T1 or independent from T1?
If T1.NAME=T2.NAME, you can automate the process with a trigger
CREATE OR ALTER TRIGGER TR_T1_AI0 FOR T1
ACTIVE AFTER INSERT POSITION 0
AS
BEGIN
INSERT INTO T2(NAME, T_ID)
VALUES (NEW.NAME, NEW.ID);
END
If T2.NAME's value is different from T1.NAME you can use a stored procedure with parameters both names:
CREATE ORA ALTER PROCEDURE XXXX(
P_NAME_T1 TYPE OF T1.NAME,
P_NAME_T2 TYPE OF T2.NAME)
AS
DECLARE VARIABLE L_ID TYPE OF T1.ID;
BEGIN
INSERT INTO T1(NAME)
VALUES (:p_NAME_T1)
RETURNING ID INTO:L_ID;
INSERT INTO T2(NAME, T_ID)
VALUES (:P_NAME_T2, :l_ID);
END
You can use both statements from the stored procedure directly in your program if it supports the returning syntax. If not, you need an additional query with SELECT NEXT VALUE FOR GENERATOR_FOR_T1 FROM RDB$DATABASE; and use the value returned from it in both INSERT statements.

Primay Key conflicts on insertion of new records

In a database application, I want to insert, update and delete records in a table of database.
Table is as below:
In this table, Ga1_ID is Primary Key.
Suppose, I insert 5 records as show currently.
In second attempt, if I want to insert 5 other records and if any of these new records contains a primary key attribute which is already present in table it show error. Its fine.
But, when I insert new 5 records... how I can verify these new records's primary key value is not present. I mean, how to match or calculate the already present primary key attributes and then insert new records.
What is the best approach to manage this sort of situation ?
use following query in dataadapter:
da=new SqlDataAdapter("select Ga1_ID from table where Ga1_ID=#pkVal",conn);
DataSet=new DataSet();
da.fill(ds);
//pass parameter for #pkVal
da.SelectCommand.Parameters(1).Value = pkValue;
if(ds.Tables[0].Rows.Count>0) //If number of rows >0 then record exists
BEGIN
messagebox.show("Primary key present");
END
Hope its helpful.
Do not check existing records in advance, i.e. do not SELECT and then INSERT. A better (and pretty common) approach is to try to INSERT and handle exceptions, in particular, catch a primary key violation if any and handle it.
Do the insert in a try/catch block, with different handling in case of a primary key violation exception and other sql exception types.
If there was no exception, then job's done, record was inserted.
If you caught a primary key violation exception, then handle it appropriately (your post does not specify what you want to do in this case, and it's completely up to you)
If you want to perform 5 inserts at once and want to make sure they all succeed or else roll back if any of them failed, then do the inserts within a transaction.
you can do a lookup first before inserting.
IF EXISTS (SELECT * FROM tableName WHERE GA1_id=#newId)
BEGIN
UPDATE tableName SET Ga1_docid = #newdocID, GA1_fieldNAme = #newName, Ga1_fieldValue = #newVal where GA1_id=#newId
END
ELSE
BEGIN
INSERT INTO tableName(GA1_ID, Ga1_docid, GA1_fieldNAme Ga1_fieldValue) VALUES (value1,val2,value3,value4)
END
If you're using SQL Server 2012, use a sequence object - CREATE SEQUENCE.
This way you can get the next value using NEXT VALUE FOR.
With an older SQL Server version, you need to create the primary key field as an IDENTITY field and use the SCOPE_IDENTITY function to get the last identity value and then increment it manually.
Normally, you would like to have a surrogate key wich is generally an identity column that will automatically increment when you are inserting rows so that you don't have to care about knowing which id already exists.
However, if you have to manually insert the id there's a few alternatives for that and knowing wich SQL database you are using would help, but in most SQL implementations, you should be able to do something like:
IF NOT EXISTS
IF NOT EXISTS(
SELECT 1
FROM your_table
WHERE Ga1_ID = 1
)
INSERT INTO ...
SELECT WHERE NOT EXISTS
INSERT INTO your_table (col_1, col_2)
SELECT col_1, col_2
FROM (
SELECT 1 AS col_1, 2 AS col_2
UNION ALL
SELECT 3, 4
) q
WHERE NOT EXISTS (
SELECT 1
FROM your_table
WHERE col_1 = q.col_1
)
For MS SQL Server, you can also look at the MERGE statement and for MySQL, you can use the INSERT IGNORE statement.

how to compare the values inside a table in sql

how to
compare the values of same table(say for eg: Order table) each and every time the record get inserted ,
if the record with same values get inserted already in same table i should not insert the new record with same values. how to do that exactly in sql server 2008
If exists(select * from Order where key_column=#some_value)
print 'data already exists'
else
Insert into Order(columns) values (#some_value,...)
I'd suggest adding a unique index on the key columns...
ALTER TABLE mytable ADD UNIQUE INDEX myindex (keycolumn1, keycolumn2, ...);
That'd make it impossible to insert a duplicate by accident.

Regarding delete a record

HI I am having a table which does not have any primary key or unique key.
How can I delete the duplicate records?
Can any one of u tell me?
The easiest way would be to copy all of the duplicates into another identical table, delete them all from the original table, then put back the duplicates (just once for each unique one of course) from the temporary table.
For example:
BEGIN TRANSACTION
CREATE TABLE Holding_Table (my_string VARCHAR(20) NOT NULL)
INSERT INTO Holding_Table (my_string)
SELECT my_string
FROM My_Table
GROUP BY my_string
HAVING COUNT(*) > 1
DELETE MT
FROM Holding_Table HT
INNER JOIN My_Table MT ON MT.my_string = HT.my_string
INSERT INTO My_Table (my_string)
SELECT my_string
FROM Holding_Table
DROP TABLE Holding_Table
COMMIT TRANSACTION
This is just a simple example with one column. You would need to adjust it for your table obviously. Then be sure to add a primary key to your table...
You would have to create a primary key first. Then you would be able to run an aggregate query and see how many duplicates there are and delete based off of the new ID. You could then remove the primary key and make another field the primary key if you so desired (or stick with the one you created).
I have done this many times when fixing ancient legacy databases.
If you use: SET ROWCOUNT 1
You can get SQL to delete only a single row, and use whatever technique you prefer to delete the identical rows one at a time.
To revert back to normal behaviour, use: SET ROWCOUNT 0
However, it would be advisable to at least add a column that allows you to uniquely identify each row so that you can avoid this problem in future. The following does the trick:
ALTER TABLE TableName ADD TableName_ID int IDENTITY NOT NULL
Now you can simply: DELETE TableName WHERE TableName_ID = ? for each of your duplicates.
Check this site on support.microsoft.com: Site
It can tell you alot of how to identify, etc.
Adding this as another answer since it's a different approach...
You could also add a new column to the table, make that one unique, and then use that to delete all but one of the duplicate rows. For example:
ALTER TABLE My_Table
ADD my_id INT IDENTITY NOT NULL
DELETE
MT1
FROM
My_Table MT1
WHERE EXISTS (
SELECT
*
FROM
My_Table MT2
WHERE
MT2.my_string = MT1.my_string AND
MT2.my_id < MT1.my_id)
ALTER TABLE My_Table
DROP COLUMN my_id