The logic has to decide whether it will insert or update the table and yet it has to check if the table is not empty since.
So the current version of the code looks like this
if ((new XYZController(mContext)).getCount() == 0) {
new XYZController(mContext).insert(object);
} else {
XYZObject obj = (new XYZController(mContext)).getByString(string);
if (obj == null) {
new XYZController(mContext).insert(object);
} else {
new XYZController(mContext).update(object);
}
}
Since I cannot do WHERE query on empty SQLite table, I have to make additional "if empty" check.
Is there a way to omit such flow and just do 1 insert and 1 update caring if the table is empty or not?
PS. If we try to make a WHERE query on empty table, the app will break on the cursor.getString()
public XYZModel getByString(String s) {
String selectQuery = "SELECT * FROM " + TABLE_XYZ
+ " WHERE " + COLUMN + "='" + s + "'";
Cursor cursor = getRow(selectQuery);
XYZModel walk = null;
if (cursor != null) {
cursor.moveToFirst();
obj = new XYZModel(cursor.getString(1), cursor.getString(2),
cursor.getString(3));
}
if (obj == null)
obj = new XYZModel();
return obj;
}
Make sure your table has reasonable uniqueness constraints, such as a PRIMARY KEY or a UNIQUE constraint on some columns
Just use an insert with a conflict resolution algorithm such as replace, for example:
db.insertWithOnConflict(tableName, null, contentValues, SQLiteDatabase.CONFLICT_REPLACE);
If the insert would violate a constraint, the conflicting rows are first deleted and then the new row is inserted. Therefore make sure you're providing a value for each column.
In case you have a database-generated INTEGER PRIMARY KEY, you can initially supply a NULL for its value (ContentValues.putNull()) and then capture the return value of the insert to get the generated primary key value.
This way you don't need the (incorrect) code that checks whether the row exists, nor do you need separate branches for inserts and updates.
Related
I am using this method to get query for this string:
public void deletedata(){
p=srt.split(",");
DatabaseHandler dba=new DatabaseHandler(this);
for(String s:p) {
dba.removeSingleproduct(s);
}
Database method is :
public boolean removeSingleproduct(String name) {
SQLiteDatabase db = this.getWritableDatabase();
return db.delete(tablename, productinserted + "=" + name, null) > 0;
}
I want to delete only one row by calling database as product inserted can have two same value.
Please help guys.
Since you're deleting with a selectedValue String,
add a single quote before and after the name
return db.delete(tablename, productinserted + " = '" + name + "'", null) > 0;
Or you can simplify your code.
public int removeSingleproduct(String name) {
return getWritableDatabase().delete(tablename, productinserted + " = ?", new String[] { name });
}
Return int - the number of rows affected if a whereClause is passed in, 0 otherwise. To remove all rows and get a count pass "1" as the whereClause.
The following will use the name to locate all rows with the provided name but only delete the first according to it's rowid (unless WITHOUT ROWID has been specified [very likely not]).
public boolean removeSingleproduct(String name) {
boolean rv = false;
SQLiteDatabase db = this.getWritableDatabase();
Cursor csr = db.query(tablename,new String[]{"rowid AS dltid"},productinserted + "=?",new String[]{name},null,null,null);
if(csr.moveToFirst()) {
rv = db.delete(tablename,"rowid=?",new String[]{Long.toString(csr.getLong(csr.getColumnIndex("dltid")))}) > 0;
}
csr.close();
return rv;
}
If you wanted to ensure that a row was only deleted if multiple rows with the same productinserted name existed, then you could simply change
if(csr.moveToFirst()) { ........
to
if(csr.moveToFirst() && csr.getCount() > 1) { .......
Note! csr.moveToLast() could be used instead of csr.moveToFirst() it would probably then delete the newest addition rather than probably deleting the oldest addition.
If you think
but I haven't defined a column called rowid
then :-
Except for WITHOUT ROWID tables, all rows within SQLite tables have a
64-bit signed integer key that uniquely identifies the row within its
table. This integer is usually called the "rowid". The rowid value can
be accessed using one of the special case-independent names "rowid",
"oid", or "rowid" in place of a column name. If a table contains a
user defined column named "rowid", "oid" or "rowid", then that name
always refers the explicitly declared column and cannot be used to
retrieve the integer rowid value.
SQL As Understood By SQLite - ROWIDs and the INTEGER PRIMARY KEY
Inserting data to SQLite table with constraint failure
I'm trying to insert data into SQLite table on Android. _id is primary key of the table and I am inserting a row using this method:
public void addSomeData(int id, String datetime) {
ContentValues contentValues = new ContentValues();
contentValues.put(KEY_ID, id);
contentValues.put(KEY_DATETIME, datetime);
mDb.insert(TABLE, null, contentValues);
}
The problem I get is that sometimes primary key constraint is validated and I would like to use something like INSERT IF NOT EXISTS, but preferably something that would work with ContentValues. What are my options? I understand that insertOrThrow() and insertWithOnConflict() methods only return different values, or should I use one of these methods?
Use insertWithOnConflict() with CONFLICT_IGNORE.
Will return ROWID/primary key of new or existing row, -1 on any error.
In my case "constraint failure" happened because of I had some tables which are depended on each other. As for the "insert if not exist", you can query with this id and you check if the cursor's count is bigger than zero. Check the method I'm already using in my app.
public boolean isRowExists(long rowId) {
Cursor cursor = database.query(this.tableName, this.columns, DBSQLiteHelper.COLUMN_ID + " = ? ", new String[] { "" + rowId }, null, null, null);
int numOfRows = cursor.getCount();
cursor.close();
return (numOfRows > 0) ? true : false;
}
to do so you could simply query the db to see if a row with that key exists and insert the new row only if the query returns no data.
This is what i am using for insert:
public long insert(String content, Date startAt, Date endAt) {
if (content == null || startAt == null) {
return 0;
}
ContentValues contentValues = new ContentValues();
contentValues.put(KEY_CONTENT, content);
contentValues.put(KEY_START_AT, startAt.getTime());
if (endAt == null) {
contentValues.putNull(KEY_END_AT);
} else {
contentValues.put(KEY_END_AT, endAt.getTime());
}
return sqLiteDatabase.insert(TABLE_NAME, null, contentValues);
}
now i want to create update method which will update last inserted row. How can i get last inserted row?
If you have an id attribute that works as a primary key, you can do a raw database query on SqlLite.
Cursor cc = this.mDb.rawQuery("SELECT *" + " FROM " + "<Your DATABASE_NAME> " +
"ORDER BY id " + "DESC LIMIT 1", null);
return cc;
Here,
1. It returns a cursor.
2. mDb is a SQLiteDatabase class instance.
3. ORDER BY id allows the query to sort by id number. As I said, if you have an id as primary key in your table, your latest entry will have the maximum id number.
4. DESC allows to sort by descending order.
5. LIMIT 1 allows to return only 1 row.
6. Always be careful when writing raw queries, white spaces inside the query can be a lot of pain when you do not handle them carefully.
For further queries you can see this tutorial. And obviously Divya's answer is also a good one.
You can use a cursor to retrieve rows and say :
cursor.moveToLast();
OR
cursor.moveToPosition(cursor.getCount() - 1);
When you insert a row in to your table the insert query returns the key of the last inserted row. You can now use this key to update this row.
for example
int newInsertedKey = sqLiteDatabase.insert(TABLE_NAME, null, contentValues);
update table_name set column_name = 'Change 2' where columnID = newInsertedKey
An efficient method would be to avoid anymore database queries to get the last updated row.
Maybe he should use something like this
public long getLastId() {
Cursor c = mDb.query(currentTableName, new String[] { "MAX(_id)" },
null, null, null, null, null, null);
try{
c.moveToFirst();
long id = c.getLong(0);
return id;
}catch(Exception e){
return 0;
}
}
where _id is column by which you identify rows
Before inserting a record in the database to validate that there is, for example I have a table that has two fields, table fields VA with Customer, First, I want to validate the field if there is no customer registration and if there insert
That i want to do in android using Sqlite
String sql = "SELECT * FROM TABLE WHERE id = '" + id + "'";
Cursor data = database.rawQuery(sql, null);
After this check by following code
if (cursor.moveToFirst()) {
// record exists
} else {
// record not found
}
Look at SQLite constraints. This really isn't Android-specific; it's a feature of SQLite.
Try this, it works for me.
String[] args = { myDataToCheck };
c = ourDatabase.query(DATABASE_TABLE, columns,
myColumnToCheck + "=?", args, null, null, null);
if (c.getCount() == 0) {
// doesn't exists
}else{
//exists
}
You have 2 possibilities :
1)make a select to check if there is the same key in database, then insert or update.
2)make directly an update. update will return the number of row updated, if the number is 0, so you can do an insert.
I am getting inconsistent results between two methods of reading the columns in an Android SQLite database.
First, this is part of a database upgrade routine as per the accepted answer here: Upgrade SQLite database from one version to another?
The technique involves moving the current table away with a temporary name, creating a new table with the new schema, and then copying relevant data from the old table into the new one before deleting the old temporary table.
The particular problem I have is when I remove a column from the schema. So, a particular column exists in the old version of the table, but not the new one.
That answer suggests using a method like this to list the columns in the table:
/**
* Returns a list of the table's column names.
*/
private List<String> getColumns(SQLiteDatabase db, final String tableName) {
List<String> ar = null;
Cursor c = null;
try {
c = db.rawQuery("SELECT * FROM " + tableName + " LIMIT 1", null);
if (c != null) {
ar = new ArrayList<String>(Arrays.asList(c.getColumnNames()));
}
} finally {
if (c != null)
c.close();
}
return ar;
}
That works fine on the old table, before I move it away with a temporary name and replace it. When I run the same query again later, on the newly-created empty table, it still lists the old table schema with the name of the column which no longer exists. It looks as if it's reusing stale cached results for that query.
If I read the columns a different way, using this instead, then it returns the new column list as expected:
private void listColumns(SQLiteDatabase db, final String tableName) {
final String query = "PRAGMA table_info(" + tableName + ");";
Cursor c = db.rawQuery(query, null);
while (c.moveToNext()) {
Log.v("MyApp", "Column: " + c.getString(1));
}
c.close();
}
The complete sequence is:
final String tempTableName = "temp_" + tableName;
table.addToDb(db); // ensure it exists to start with
// get column names of existing table
final List<String> columns = getColumns(db, tableName);
// backup table
db.execSQL("ALTER TABLE " + tableName + " RENAME TO " + tempTableName);
// create new table
table.addToDb(db);
// delete old columns which aren't in the new schema
columns.retainAll(getColumns(db, tableName));
// restore data from old into new table
String columnList = TextUtils.join(",", columns);
db.execSQL(String.format("INSERT INTO %s (%s) SELECT %s from %s", tableName, columnList, columnList,
tempTableName));
// remove backup
db.execSQL(DROP_TABLE + tempTableName);
What's the reason for the different results?
I assume you have done something similar to this:
ALTER TABLE "main"."mytable" RENAME TO "newtable";
CREATE TABLE "main"."mytable" ("key1" text PRIMARY KEY,"key2" text,"key3" text);
INSERT INTO "main"."mytable" SELECT "key1","key2","key3" FROM "main"."newtable";
DROP TABLE "main"."newtable";
If you have, please share the equivalent code, just to rule out any errors with this part.
I never got to the bottom of this. I just ended up using the second method I mentioned, which doesn't exhibit the problem.