Here is the code that works:
Connection c = ds.getConnection();
c.setAutoCommit(false);
PreparedStatement stmt = c.prepareStatement("INSERT INTO items (name, description) VALUES(?, ?)");
while (!(items = bus.take()).isEmpty()) {
for (Item item : items) {
stmt.setString(1, item.name);
stmt.setString(2, item.description);
stmt.addBatch();
}
stmt.executeBatch();
c.commit();
}
But now I need to populate another table where id is a foreign key.
If I use INSERT with RETURNING id then executeBatch fails with "A result was returned when none was expected" error.
I see several ways to solve this
Do individual insert rather than the batch insert.
Replace serial id with client generated guid.
Use some kind of a stored procedure to perform the batch insert and return a list of ids.
Of the three methods that I see the last one seems to preserve both the efficiency of batch insert and return the ids, but it is also the most complex for me as I have never written stored procedures.
Is there a better way to batch insert and get the IDs? I have no problem using postgresql specific API rather than jdbc.
If not, could any one sketch such a stored procedure?
Here is the table schema:
CREATE UNLOGGED TABLE items
(
id serial,
name character varying(1000),
description character varying(10000)
)
WITH (
OIDS=FALSE
);
Something like this should work:
// tell the driver you want the generated keys
stmt = c.prepareStatement("INSERT ... ", Statement.RETURN_GENERATED_KEYS);
stmt.executeBatch();
// now retrieve the generated keys
ResultSet rs = stmt.getGeneratedKeys();
while (rs.next()) {
int id = rs.getInt(1);
.. save the id somewhere or update the items list
}
I think (I am not sure!) that the keys are returned in the order they were generated. So the first row from the ResultSet should map to the first "item" from the list you are processing. But do verify that!
Edit
If that doesn't work, try specifying the actual columns for which the values are generated:
stmt = c.prepareStatement("INSERT ... ", new String[] {"id"});
Related
String sql = "Select MAX(ORDERLINEID) From ORDERLINESTABLE";
ResultSet rst;
rst = stmt.executeQuery(sql);
if(rst.next())
{
next = rst.getInt("ORDERLINEID");
next++;
}
I have a table called ORDERLINESTABLE in my database which is currently empty. I have run the above code with the aim is to get the highest integer stored in the ORDERLINEID column allowing me to increment it when adding items to the database.
I expected this query to return nothing as the table is empty but when debugging I noticed that the search is returning true for the rst.next() method.
Does anyone have any idea why this would be? I have looked at the resultset.next() documentation and as far as I can see it should return false.
When in doubt, look at your data. Here is a sample query from any db engine.
select max(field) maxValue
from table
where 1=3
It will yield
maxValue
Null
In other words, your query is returning one record with a value of null.
It is much better to fetch the ORDERLINEID filled in by the database after the INSERT statement. Make the column ORDERLINEID of type INT AUTOINCREMENT.
String sql = "INSERT INTO ORDERLINESTABLE(xxx, yyy) VALUES (?, ?)";
try (PreparedStatement stmt = conn.prepareStatement(sql, Statement.RETURN_GENERATED_KEYS)) {
stmt.setString(1, xxx);
stmt.setInt(2, yyy);
int updateCount = stmt.executeUpdate(); // 1
try (ResultSet id = stmt.getGeneratedKeys()) {
if (id.next()) { // 'if' as just 1 row inserted.
int orderLineId = id.getInt(1); // 1 key per row.
}
}
}
Java has a database independent way to fetch the generated keys of an INSERT. That is a lot safer than taking the MAX afterwards or before, in a multi-user environment.
Scenarios for wrong IDs are numerous in a multiuser environment:
first SELECT
second SELECT
second increment for new ID
first increment for new ID
first INSERT
second INSERT
I have an assignment where I need to update records using a PreparedStatement. Once the record have been updated as we know update query return count, i.e., number of row affected.
However, instead of the count I want the rows that were affected by update query in response, or at least a list of id values for the rows that were affected.
This my update query.
UPDATE User_Information uInfo SET address = uInfo.contact_number || uInfo.address where uInfo.user_id between ? AND ?;
Normally it will return count of row affected but in my case query should return the ids of row or all the row affected.
I have used the returning function of PostgreSQL it is working but is not useful for me in that case.
i have used returning function of PostgreSQL but is not useful for me
It should be. Perhaps you were just using it wrong. This code works for me:
sql = "UPDATE table1 SET customer = customer || 'X' WHERE customer LIKE 'ba%' RETURNING id";
try (PreparedStatement s = conn.prepareStatement(sql)) {
s.execute(); // perform the UPDATE
try (ResultSet rs = s.getResultSet()) {
// loop through rows from the RETURNING clause
while (rs.next()) {
System.out.println(rs.getInt("id")); // print the "id" value of the updated row
}
}
}
The documentation indicates that we can also use RETURNING * if we want the ResultSet to include the entire updated row.
Update:
As #CraigRinger suggests in his comment, the PostgreSQL JDBC driver does actually support .getGeneratedKeys() for UPDATE statements too, so this code worked for me as well:
sql = "UPDATE table1 SET customer = customer || 'X' WHERE customer LIKE 'ba%'";
try (PreparedStatement s = conn.prepareStatement(sql, Statement.RETURN_GENERATED_KEYS)) {
s.execute(); // perform the UPDATE
try (ResultSet rs = s.getGeneratedKeys()) {
while (rs.next()) {
System.out.println(rs.getInt(1)); // print the "id" value of the updated row
}
}
}
Thanks, Craig!
You might be able to use JDBC's support for getting generated keys. See the Connection.prepareStatement(String sql, int[] columnIndexes) API method, then use Statement.getGeneratedKeys() to access the results.
The spec says "the driver will ignore the array if the SQL statement is not an INSERT statement" but I think PostgreSQL's JDBC driver will actually honour your request with other statement types too.
e.g.
PreparedStatement s = conn.prepareStatement(sql, new String[] {'id'})
s.executeUpdate();
ResultSet rs = s.getGeneratedKeys();
Otherwise, use RETURNING, as Gord Thompson describes.
There are two way of doing it
1. by passing an array of column name or index of column prepareStatement
i.e conn.prepareStatement(sql, new String[] {'id','uname'})
and
2. by using Statement.RETURN_GENERATED_KEYS in prepareStatement.
My code is for this i.e as per my requirement i have developed my code you can have a look for better idea.
private static final String UPDATE_USER_QUERY= "UPDATE User_Information uInfo SET address = uInfo.contact_number || uInfo.address where uInfo.user_id between ? AND ?;";
//pst = connection.prepareStatement(UPDATE_USER_QUERY,columnNames);
pst = connection.prepareStatement(UPDATE_USER_QUERY,Statement.RETURN_GENERATED_KEYS);
ResultSet rst = pst.getGeneratedKeys();
List<UserInformation> userInformationList = new ArrayList<UserInformation>();
UserInformation userInformation;
while (rst.next()){
userInformation = new UserInformation();
userInformation.setUserId(rst.getLong("user_id"));
userInformation.setUserName(rst.getString("user_name"));
userInformation.setUserLName(rst.getString("user_lName"));
userInformation.setAddress(rst.getString("address"));
userInformation.setContactNumber(rst.getLong("contact_number"));
userInformationList.add(userInformation);
}
That think i need to achieve in this case.
Hope so this will help you a lot.
My source code has the following structure:
SourceFolder
AddProduct.jsp
Source Packages
-Controller(Servlets)
SaveProduct.java
-Model(Db Operations)
ProductDbOperations.java
I am inserting a new product into the product table and at the same time I am inserting an entry into product_collection table (product_id | collection_id).
To insert an entry into the product_collection table i need to get generated id from product table. After that a new entry is inserted into the product_collection table.
Also, I am not using any Framework and am using Netbeans 7.3.
Problem:
A new entry is inserted into the product table with this piece of code
IN: ProductDbOperations.java
try
{
this.initConnection(); // Db connection
pst = cn.prepareStatement("INSERT INTO product values('"+name+"', "+quantity+", "+price+")");
rs = pst.executeUpdate();
}
catch(SQLException ex)
{
}
I Also used the solution at following link which doesn't works for me.
I didn't got any SQL exception
How to get the insert ID in JDBC?
so help me find out why this code not working for me .
Thanks a million.
Not all drivers support the version of getGeneratedKeys() as shown in the linked answer. But when preparing the statement, you can also pass the list of columns that should be returned instead of the "flag" Statement.RETURN_GENERATED_KEYS (and passing the column names works more reliably in my experience)
Additionally: as javaBeginner pointed out correctly, your usage of a prepared statement is wrong. The way you do it, will still leave you wide open to SQL injection.
// run the INSERT
String sql = "INSERT INTO product values(?,?,?)";
pst = cn.prepareStatement(sql, new String[] {"PRODUCT_ID"} );
pst.setString(1, name);
pst.setInt(2, quantity);
pst.setInt(3, price);
pst.executeUpdate();
// now get the ID:
ResultSet rs = pst.getGeneratedKeys();
if (rs.next()) {
long productId = rs.getLong(1);
}
Note that the column name passed to the call is case-sensitive. For Oracle the column names are usually uppercase. If you are using e.g. Postgres you would most probably need to pass new String[] {"product_id"}
The way you are using is not the proper way of using preparedstatement
use the following way
pst = cn.prepareStatement("INSERT INTO product values(?,?,?)");
pst.setString(1,name);
pst.setInt(2,quantity);
pst.setInt(3,price);
pst.executeUpdate();
Yes there is a way to retrieve the key inserted by SQL. You can do it by:
Using Statement.RETURN_GENERATED_KEYS in your previous insert and get the key which can be used in further insert
e.g:
String query = "INSERT INTO Table (Col2, Col3) VALUES ('S', 50)";
Statement stmt = con.createStatement();
int count = stmt.executeUpdate(query, Statement.RETURN_GENERATED_KEYS);
I am using oracle sequence for inserting log id into tableA as follows,
String SQL_PREP_INSERT = "INSERT INTO tableA (LOG_ID,USER_ID,EXEC_TIME) VALUES"
+ " (logid_seq.nextval, ?, ?)";
Then getting the recently inserted value,
String SQL_PREP_SEL = "SELECT max(LOG_ID) FROM tableA ";
stmt = con.prepareStatement(SQL_PREP_SEL);
stmt.execute();
ResultSet rs = stmt.getResultSet();
if (rs.next()) {
logid = rs.getInt(1);
}
And inserting it into tableB,
String SQL_PREP_INSERT_DETAIL = "INSERT INTO tableB (LOG_ID, RESPONSE_CODE, RESPONSE_MSG) VALUES"
+ " (?, ?)";
stmt = con.prepareStatement(SQL_PREP_INSERT_DETAIL);
stmt.setInt(1, logid);
stmt.setString(2, respCode);
stmt.setString(3, respMsg);
stmt.execute();
Is there a way to generate sequence in Java instead of Oracle and insert into both tables at once, instead of selecting from tableA and inserting into tableB?
In general, selecting the MAX(log_id) is not going to give you the same value that logid_seq.nextval provided. Assuming that this is a multi-user system, some other user could have inserted another row with a larger log_id value than the row you just inserted before your query is executed.
Assuming that both INSERT statements are run in the same session, the simplest option is probably to use the logid_seq.currval in the second INSERT statement. currval will return the last value of the sequence that was returned to the current session so it will always return the same value that was generated by the nextval call in the first statement.
INSERT INTO tableB (LOG_ID, RESPONSE_CODE, RESPONSE_MSG)
VALUES( logid_seq.currval, ?, ? )
Alternatively, you could use the RETURNING clause in your first statement to fetch the sequence value into a local variable and use that in the second INSERT statement. But that is probably more work than simply using the currval.
String QUERY = "INSERT INTO students "+
" VALUES (student_seq.NEXTVAL,"+
" 'Harry', 'harry#hogwarts.edu', '31-July-1980')";
// load oracle driver
Class.forName("oracle.jdbc.driver.OracleDriver");
// get database connection from connection string
Connection connection = DriverManager.getConnection(
"jdbc:oracle:thin:#localhost:1521:sample", "scott", "tiger");
// prepare statement to execute insert query
// note the 2nd argument passed to prepareStatement() method
// pass name of primary key column, in this case student_id is
// generated from sequence
PreparedStatement ps = connection.prepareStatement(QUERY,
new String[] { "student_id" });
// local variable to hold auto generated student id
Long studentId = null;
// execute the insert statement, if success get the primary key value
if (ps.executeUpdate() > 0) {
// getGeneratedKeys() returns result set of keys that were auto
// generated
// in our case student_id column
ResultSet generatedKeys = ps.getGeneratedKeys();
// if resultset has data, get the primary key value
// of last inserted record
if (null != generatedKeys && generatedKeys.next()) {
// voila! we got student id which was generated from sequence
studentId = generatedKeys.getLong(1);
}
}
is there any way of retrieving database generated values in Java other than IDENTITY fields? I can easily get IDENTITY values from a ResultSet, but I'd like to get the value of a date field which has been generated by the database (CURRENT_TIMESTAMP). I prefer not to send another SELECT query to get the date.
statement = connection.prepareStatement("INSERT INTO foo (bar_date) VALUES (CURRENT_TIMESTAMP)");
ResultSet generatedKey = statement.getGeneratedKeys();
while (generatedKey.next()) {
// read the key..., this unfortunately only returns IDENTITY columns.
}
Have you tried something like
INSERT INTO foo (bar_date)
OUTPUT INSERTED.bar_date
VALUES (CURRENT_TIMESTAMP)
Have a look at OUTPUT Clause (Transact-SQL)
Assuming you're using SQL 2005 or later, you could add an OUTPUT clause to your insert statement to have it return a result set.
I'm not sure of the Java syntax to retrieve the resultset, but it would be something like the following:
statement = connection.prepareStatement("INSERT INTO foo OUTPUT inserted.id, inserted.bar_date (bar_date) VALUES (CURRENT_TIMESTAMP)"); //guessing the name of the id column
ResultSet returnSet = statement.execute(); // or however you do this
while (resultset.next()) {
// resultset will have two columns - id and bar_date
}