TL;DR: I want to create a pure function that will return a sequence of insert operations closed inside a single DBIOAction object (e.g def foo(): DBIOAction). This is trivial when IDs are NOT auto-incremented, it gets difficult for me when ID's ARE auto-incremented by DB).
I've been using Slick for more then a week now and what I really like about it is ability to close sequence of operations as a DBIOAction, that can be returned by my function and later can be applied to DB as a side effect.
That worked all fine when I was inserting data to a single table. But how should I deal with situation when an ID of one inserted row must be used in a second row that I'm inserting (like any 1-many relation), given the fact that IDs are auto-incremented?
Imagine we have a CATEGORIES table with columns: id, name, parent_id, where parent_id is either NULL or points to other row within the table.
If the IDs were not auto-incremented, this is trivial:
def insert(cId: Int, cName: String, subcId: Int, subcName: String) =
DBIO.seq(
categories +=(subcId, subcName, None),
categories +=(cId, cName, Some(subcId))
)
But when ID ARE auto-incrememnted, this get's tricky. As stated in the documentation, I need to use returning combined with projection to column that needs to be returned, like:
val categoryId =
(categories returning categories.map(_.id)) += (0, name, None)
Now I can use returned ID when inserting the second row. Awesome. But then how can I close sequences of DB actions, since I am returned with the ID, but not the action itself?
def insert(cName: String, subcName: String): DBIOAction[Int, Stream, Write] = ???
So I will answer my own question as somebody might find it weird at first as I did.
val categoryId =
(categories returning categories.map(_.id)) += (0, name, None)
Trick is that categoryId is also a DBIOAction, not an Int as I originally expected it to be. Trivial, but there you go.
This is how I would do (I've done something similar in a project of mine):
def insert(cName: String, subcName: String) = {
for {
id <- (categories returning categories.map(_.id)) += (subcName, None)
_ <- categories +=(cName, Some(id))
} yield ()
and then I'd call the function like:
db.run(insert(categoryName, subCategoryName))
Related
I have a database from which data is coming into a QTableWidget. The table in the database has the following Columns,
ID (Primary key, auto-increment value)
Name
Location
The QTableWidget has the following columns (that I have added)
ID (this column, I have hidden. and it contains the value of "ID" column from the Database Table)
Sr # (Represents the Row Number of the table)
Name (Contains "name" from the database table)
Location (Contains "Location from the database table)
Actions (Contains a Delete Button for that Row)
By hidden, I mean to say that I have made this column hidden using the folliwng command,
self.ui.table.setColumnHidden(0, True);
This is how I am populating my QTableWidget and creating a Delete Function,
def get_data(self):
mycursor = self.DB.cursor()
Subquery = "select id, name, location "
Subquery += " from tbl_person"
mycursor.execute(Subquery)
numcols = len(mycursor.fetchall()[0])
mycursor.execute(Subquery)
numrows = len(mycursor.fetchall())
self.ui.table.setRowCount(numrows)
self.ui.table.setColumnCount(numcols+2)
mycursor.execute(Subquery)
tablerow = 0
for row in mycursor.fetchall():
layout = QHBoxLayout()
layout.setContentsMargins(0, 0, 0, 0)
layout.setSpacing(0)
delete_button = QPushButton("Delete Data")
delete_button.clicked.connect(self.executeDeleteFunction)
# delete_button.setStyleSheet(delete_push_button) -> Only for styling
self.ui.table.setItem(tablerow, 0, PySide2.QtWidgets.QTableWidgetItem(str(row[0])))
self.ui.table.setItem(tablerow, 1, PySide2.QtWidgets.QTableWidgetItem(str(tablerow+1)))
self.ui.table.setItem(tablerow, 2, PySide2.QtWidgets.QTableWidgetItem(str(row[1])))
self.ui.table.setItem(tablerow, 3, PySide2.QtWidgets.QTableWidgetItem(str(row[2])))
self.ui.table.setCellWidget(tablerow, 4, delete_button)
tablerow += 1
self.ui.table.setColumnHidden(0, True)
#self.ui.registered_table.horizontalHeader().setSectionResizeMode(PySide2.QtWidgets.QHeaderView.Stretch)
self.ui.table.resizeColumnsToContents()
def executeDeleteFunction(self):
self.person_id = self.ui.table.selectionModel().selectedIndexes()[0]
self.person_id = self.person_id.row()
mycursor = self.DB.cursor()
sql = "delete from tbl_person where id = %s"
val = (id, )
mycursor.execute(sql, val)
print("Deletion Successful")
On the Deletion Function, what this code does is basically gets the value of the **Sr # ** Column from the QTableWidget and deletes the data according to that, i.e. it is getting me the value from the visible first column and not the actual first column. But, I want the data from the "ID" column of the QTableWidget which is hidden
I tried to look up on how to get the value from the first hidden column on the QTableWidget and ended up with this link: How to get data from hidden 'id' column in QtableWidget
This apparently solves my issue but I can not seem to make it work for my code. I don't want to retrieve values of multiple Rows but only of one row so how do I do this (as I am only deleting one row. But in the question mentioned, I believe that it is getting data from multiple rows due to that for each loop)?
Moreover, I tried to find help regarding the functionality of sibling function (which is provided in the answer of above question) however I could not find any good resource on this function (i.e. how to use this, or some practical example and etc.)
I tried the following with Sibling function to obtain the value of first hidden column of the Selected Row but it did not work,
self.value = self.table.selectedItems()[0]
self.value = sibling(self.value.row(), 0)
There are some conceptual problems with the given code.
First of all, the QtSql module should be preferred instead of artificially creating a model. For basic tables, QSqlTableModel is fine enough, while for custom queries, QSqlQueryModel is a good choice.
Now the problem is that UI-based selection is always based on visible items: if you select a row in a view that has hidden columns, you will not get the hidden indexes that belong to those columns.
In order to get the indexes (as in QModelIndex) of hidden columns on a table widget, the only way is the same for a table view: you need to access the model and get the index for the row, or you get the actual model index and then get the sibling (which is conceptually the same, as the underlying function does):
item = self.table.selectedItems()[0]
index = self.table.indexForItem(index)
firstRowIndex = index.sibling(index.row(), 0)
sqlIndex = firstRowIndex.data() # might be a string
Note that you can also use siblingAtColumn():
firstRowIndex = index.siblingAtColumn(0)
That's because when you create QTableWidget items, you're actually creating a new model, and the row for that model doesn't reflect the actual "row" of that index in the source model; items in the second row will return 1 for row(), even if their actual row is different, and that's because that item has been added as second to the table widget, since it's the second item in the query.
So, the solution is that you either get the incremental row value for the first column index sibling, or you use one of the predefined Sql models.
For simple models, the latter solution is fine enough, but if you need more complex models, the first is certainly more accurate and reliable.
like the title mentions I want a Query that gets a random row of the existing database. After that I want to save the data which is in a specific column of that row in a variable for further purposes.
The query I have at the moment is as follows:
#Query("SELECT * FROM data_table ORDER BY RANDOM() LIMIT 1")
fun getRandomRow()
For now I am not sure if this query even works, but how would I go about writing my function to pass a specific column of that randomly selected row to a variable?
Ty for your advice, tips and/or solutions!
Your query is almost correct; however, you should specify a return type in the function signature. For example, if the records in the data_table table are mapped using a data class called DataEntry, then the query could read as shown below (note I've also added the suspend modifier so the query must be run using a coroutine):
#Query("SELECT * FROM data_table ORDER BY RANDOM() LIMIT 1")
suspend fun getRandomRow(): DataEntry?
If your application interacts with the database via a repository and view model (as described here: https://developer.android.com/topic/libraries/architecture/livedata) then the relevant methods would be along the lines of:
DataRepository
suspend fun findRandomEntry(): DataEntry? = dataEntryDao.getRandomRow()
DataViewModel
fun getRandomRecord() = viewModelScope.launch(Dispatchers.IO) {
val entry: DataEntry? = dataRepository.findRandomEntry()
entry?.let {
// You could assign a field of the DataEntry record to a variable here
// e.g. val name = entry.name
}
}
The above code uses the view model's coroutine scope to query the database via the repository and retrieve a random DataEntry record. Providing the returning DataEntry record is not null (i.e. your database contains data) then you could assign the fields of the DataEntry object to variables in the let block of the getRandomRecord() method.
As a final point, if it's only one field that you need, you could specify this in the database query. For example, imagine the DataEntry data class has a String field called name. You could retrieve this bit of information only and ignore the other fields by restructuring your query as follows:
#Query("SELECT name FROM data_table ORDER BY RANDOM() LIMIT 1")
suspend fun getRandomRow(): String?
If you go for the above option, remember to refactor your repository and view model to expect a String instead of a DataEntry object.
It's rare that I find an issue that hasn't already be answered but I've been searching for this for 3 days and haven't found anything yet.
I'm aiming to create a page for inputing records in a 'spreadsheet' like format. I've used inline editing in SQLFORM.grid from this slice.
The problem I'm having is that when one of the fields is a reference to another table, the row being use in the lambda function is taking the row of the reference table rather than the row in the grid.
Here is an example:
Model
db.define_table('people',
Field('name', 'string'),
format = '%(name)s',
)
db.define_table('animals',
Field('name', 'string'),
Field('pet_owner', 'reference people'),
format = '%(name)s',
)
Controller
def index():
#process submitted form
if len(request.post_vars) > 0:
print request.post_vars
for key, value in request.post_vars.iteritems():
(field_name,sep,row_id) = key.partition('_row_')
if row_id:
db(db.animals.id == row_id).update(**{field_name:value})
db.animals.name.represent = lambda value,row: SQLFORM.widgets.string.widget(db.animals.pet_owner,value, **{'_name':'name_row_%s' % row.id})
db.animals.pet_owner.represent = lambda value,row: SQLFORM.widgets.options.widget(db.animals.pet_owner,value, **{'_name':'pet_owner_row_%s' % row.id})
grid = SQLFORM.grid(db.animals,
selectable= lambda ids : redirect(URL('animals',vars=request._get_vars)),
)
grid.elements(_type='checkbox',_name='records',replace=None) #remove selectable's checkboxes
return dict(grid=grid)
At first it appears that the grid is working correctly. However, when inspecting the drop-downs for the reference fields, if two consecutive rows have the same value (e.g. two animals with the same owner) the same row is used in the name (pet_owner_row_1) which means that the value being passed to process the submitted form is not an integer (e.g. 3) but as the integers separated by pipes (e.g. '|3|3|').
I've confirmed that this is where the issue is by changing the represent to
db.animals.pet_owner.represent = lambda value,row: row
which shows the exact same row data for different animals.
Here is an image showing the inspector on the form items: http://i.stack.imgur.com/oFtF8.png
How can I get the row id of the grid's row rather than the id of the reference?
Any help is greatly appreciated!
This is a bug that has recently been fixed in the master branch but not released yet. In the meantime, a workaround is to temporarily change the field to an integer type:
db.animals.pet_owner.type = 'integer'
db.animals.pet_owner.represent = lambda value,row: SQLFORM.widgets.options.widget(
db.animals.pet_owner,value, **{'_name':'pet_owner_row_%s' % row.id})
I used a code, generated from slick code generator.
My table has more than 22 columns, hence it uses HList
It generates 1 type and 1 function:
type AccountRow
def AccountRow(uuid: java.util.UUID, providerid: String, email: Option[String], ...):AccountRow
How do I write compiled insert code from generated code?
I tried this:
val insertAccountQueryCompiled = {
def q(uuid:Rep[UUID], providerId:Rep[String], email:Rep[Option[String]], ...) = Account += AccountRow(uuid, providerId, email, ...)
Compiled(q _)
}
I need to convert Rep[T] to T for AccountRow function to work. How do I do that?
Thank you
;TLDR; Not possible
Explanation
There are two levels of abstraction in Slick: Querys and DBIOActions.
When you're dealing with Querys, you have to access your schema definitions, and rows, Reps and, basically, it's very constrained as it's the closest level of abstraction to the actual DB you're using. A Rep refers to an hypothetical value in the database, not in your program.
Then you have DBIOActions, which are the next level... not just some definition of a query, but the execution of it. You usually get DBIOActions when getting information out of a query, like with the result method or (TADAN!) when inserting rows.
Inserts and Updates are not queries and so what you're trying to do is not possible. You're dealing with DBIOAction (the += method), and Query stuff (the Rep types). The only way to get a Rep inside a DBIOAction is by executing a Query and obtaining a DBIOAction and then composing both Actions using flatMap or for comprehensions (which is the same).
I'd like to be able to update a previously persisted object for which I have an id without having to retrieve it first. The main thing that I'm trying to avoid is having to copy multiple values into the object's fields when that object has been retrieved from the database. I have these values in a map with keys corresponding to the field names so it's trivial to create the object via a constructor with the map as an argument. Unfortunately, an object created this way results in a new database record when saved even though the id field is set to that of an existing record.
I'm currently using a slight variation on one of the examples shown here for copying Groovy class properties but it's not a very elegant solution for multiple reasons.
Basically I'd like to be able to do something like this:
class Foo {
int a
String b
}
def data = [id: 99, a: 11, b: "bar"] //99 is the id of an existing record
def foo = new Foo(data)
foo.update() //or some other comparable persistence mechanism
Thanks
As long as your map keys have the same name as your object properties, you can use executeUpdate without specifying the individual property names with a closure or function like the following:
def updateString = { obj, map ->
def str = ""
map.each { key, value ->
str += "${obj}.${key}=:${key},"
}
return str[0..-2]
}
def data= [foo:"bar", machoMan:"RandySavage"]
In this case, println updateString("f", data) returns "f.foo=:foo,f.machoMan=:machoMan".
Then you can do this:
Foo.executeUpdate("update Foo f set ${updateString("f", data)}", data)
Or of course you could combine that all together into one closure or function.
You can use the executeUpdate method on the GORM domain class:
Foo.executeUpdate("update Foo f set f.a=:a, f.b=:b where f.id=:id", data)