I have a SQL table of tags like this:
Id | Tag
-----------------
1 | car
1 | red
1 | sport
2 | car
2 | red
2 | SUV
And I want to retrieve ONLY the ID for exact matching search strings. So, using LINQ, I'd like to query for:
"car,red"
and have it return: 1 and 2.
Then searching for "car,red,sport" would return 1 only.
I'm not sure at all how to do this using LINQ. If I do the following (using EF context and table as example):
string[] tags = {"car","red","sport"}
var query = context.CarTags.Where(a => tags.Contains(a.Tag)).Select(s=>s);
...will of course return both 1 & 2.
So, how do I do this using LINQ?
string[] tags = {"car","red","sport"}
var query = context.CarTags
.GroupBy(a => a.Id)
.Where(g => tags.All(t => g.Any(a => a.Tag == t)))
.Select(g => g.Key);
I'm not sure if it's gonna work with LINQ to Entites, but would definitely give it a try.
All methods used are listed as supported within Supported and Unsupported LINQ Methods (LINQ to Entities) list, so should work. But even if it is, I would probably check the generated SQL to make sure it's not over-complicated and not very effective.
Related
I am crafting a sql query that dynamically builds a where clause. I was able to transform the separate pieces of the where clause as return rows like so:
-------------------------------------------
| ID | Query Part |
-------------------------------------------
| TOKEN 1 | (A = 1 OR B = 2) |
-------------------------------------------
| TOKEN 2 | ([TOKEN 1] or C = 3 |
-------------------------------------------
| TOKEN 3 | ([TOKEN 2] and D = 4) |
-------------------------------------------
My goal is to wrap the current return results above in a stuff and or replace (or something entirely different I hadn't considered) to output the following result:
(((A=1 OR B=2) OR C=3) AND D=4)
Ideally there would be no temp table necessary but I am open to recommendations.
Thank you for any guidance, this has had me pretty stumped at work.
Its unusual. It looks like the query part you want is only Token 3. Then the process should replace any [token] tags in this query part with the corresponding query parts. With the subsequent resulting query part, again the process should replace any [token] tags with the corresponding query parts. This continues until there are no more [token] tags to replace.
I think there should there be a way of indicating the master query (ie token 3) , then use a recursive common table expression to build the expression up until there are no more [token]s.
How can I implement an SQL query like this, in SparkSQL 2.0 using DataFrames and Scala language? I've read a lot of posts but none of them seems to achieve what I need, or if you can point me one, would do. Here's the problem:
UPDATE table SET value = 100 WHERE id = 2
UPDATE table SET value = 70 WHERE id = 4
.....
Suppose that you have a table table with two columns like this:
id | value
--- | ---
1 | 1
2 | null
3 | 3
4 | null
5 | 5
Is there a way to implement the above query, using map, match cases, UDFs or if-else statements? The values that I need to store in the value field are not sequential, so I have specific values to put there. I'm aware too that it is not possible to modify a immutable data when dealing with DataFrames. I have no code to share because I can't get it to work nor reproduce any errors.
Yes you can, it's very simple. You can use when and otherwise.
val pf = df.select($"id", when($"id" === 2, lit(100)).otherwise(when($"id" === 4, lit(70)).otherwise($"value")).as("value"))
Situation
I trying to set up a database schema to store translations, between different languages. So far it looks like this (simplyfied):
class Language(models.Model):
tag = models.CharField(max_length=2)
def __unicode__(self):
return self.tag
class Phrase(models.Model):
name = models.TextField()
language = models.ForeignKey(Language)
def __unicode__(self):
return self.name
class Meta:
unique_together = ("name", "language")
index_together = [
["name", "language"]
]
class Translation(models.Model):
phrase1 = models.ForeignKey(Phrase, related_name="translation_as_1")
phrase2 = models.ForeignKey(Phrase, related_name="translation_as_2")
def __unicode__(self):
return self.phrase1.name + " <=> " + self.phrase2.name
class Meta:
unique_together = ("phrase1", "phrase2")
index_together = [
["phrase1", "phrase2"]
]
This database schema seems logical to me. I store phrases in different languages and then have translations that contain exactly two phrases.
Problem
The problem is, that the queries, that result out of this schema, look kind of nasty. For instance:
from django.db.models import Q
name = "my phrase"
translations = Translation.objects.filter(Q(phrase1__name=text)|Q(phrase2__name=text))
translated_names = []
for translation in translations:
name1 = translation.phrase1.name
name2 = translation.phrase2.name
if name1 == name:
translated_names.append(name2)
else:
translated_names.append(name1)
I always have to include the "OR" relationship, to make sure, that I get all the possible translations, since the phrase could be stored as phrase1 or phrase2. On top of that, I have to filter my result afterwards to get the correct translated_name (for loop).
Further Explaination
Before I switched to the described schema, I had the following schema instead (Phrase and Language are the same as before):
class Translation(models.Model):
phrase = models.ForeignKey(Phrase)
name = models.TextField()
def __unicode__(self):
return self.phrase.name + " => " + self.name
class Meta:
unique_together = ("phrase", "name")
index_together = [
["phrase", "name"]
This schema let me make queries like this:
from django.db.models import Q
name = "my phrase"
translations = Translation.objects.filter(phrase__name=text)
translated_names = [t.name for t in translations]
This looks much nicer, and is of course faster. But this schema had the disadvantage, that it presents translations only in one direction, so I moved to the other one, which isn't quite what I want as well, because too slow and too complicated queries.
Question
So is there a good schema for this kind of problem, that I maybe overlook?
Remark
I'm not only interested in Django related answers. A pure SQL schema for this kind of problem would also be interesting for me.
This is the way that I have done it in the past. Adapt it for your naming convention.
Suppose that I had a table with a name and other columns in it like this
Table TR_CLT_clothing_type
clt_id | clt_name | other columns ....
--------------------------------------
1 | T Shirt ...
2 | Pants ...
Now if I decided that it needs translations, first I make a languages table
Table TR_LNG_language
lng_id | lng_name | lng_display
-------------------------------
1 | English | English (NZ)
2 | German | Deutsch
I also need to store the current language in the database (you will see why soon). It will only have one row
Table TA_INF_info
inf_current_lng
---------------
1
Then I drop the clt_name column from my clothing table TR_CLT_clothing_type. Instead I make relation table.
Table TL_CLT_clothing_type
clt_id | lng_id | clt_name
--------------------------
1 | 1 | T Shirt
1 | 2 | (German for T-Shirt)
2 | 1 | Pants
2 | 2 | keuchen (thank you google translate)
Now to get the name, you want to make a stored procedure for it. I have not attempted this in ORM.
CREATE PROCEDURE PS_CLT
#clt_id int
AS
SELECT lng.clt_name, clt.*
FROM TR_CLT_clothing_type clt
JOIN TL_CLT_clothing_type lng
ON lng.clt_id = clt.clt_id
WHERE clt.clt_id = #clt_id AND
lng.lng_id in (SELECT inf_current_lng FROM TA_INF_info)
This stored proc will return the name in the current language and all other columns for a specified language. To set the language, set the clt_current_lng in the TA_INF_info table.
Disclaimer: I don't have anything to check the syntax of what I have typed but it should hopefully be straightforward.
-- EDIT
There was a concern to be able to do "give me all translations for word X in language Y to language Z"
There is a "not so elegant" way to do this with the schema. You can do something like
for each table in database like "TL_%"
SELECT name
FROM table
WHERE id IN ( SELECT id
FROM table
WHERE name = #name
AND lng_id = german
)
AND lng_id = english
Now I would imagine that this would require some auto-generated SQL code but I could pull it off.
I have no idea how you would do this in ORM
i have string of value like "4,3,8"
and i had comma separated column in table as below.
ID | PrdID | cntrlIDs
1 | 1 | 4,8
2 | 2 | 3
3 | 3 | 3,4
4 | 4 | 5,6
5 | 5 | 10,14,18
i want only those records from above table which match in above mention string
eg.
1,2,3 this records will need in output because its match with the passing string of "4,3,8"
Note : i need this in entity framework LINQ Query.
string[] arrSearchFilter = "4,3,8".Split(',');
var query = (from prdtbl in ProductTables
where prdtbl.cntrlIDs.Split(',').Any(x=> arrSearchFilter.Contains(x))
but its not working and i got below error
LINQ to Entities does not recognize the method 'System.String[] Split(Char[])' method, and this method cannot be translated into a store expression.
LINQ to Entities tries to convert query expressions to SQL. String.Split is not one of the supported methods. See http://msdn.microsoft.com/en-us/library/vstudio/bb738534(v=vs.100).aspx
Assuming you are unable to redesign the database structure, you have to bypass the SQL filter and obtain ALL records and then apply the filter. You can do this by using ProductTables.ToList() and then using this in second query with the string split, e.g.
string[] arrSearchFilter = "4,3,8".Split(',');
var products = ProductTables.ToList();
var query = (from prdtbl in products
where prdtbl.cntrlIDs.Split(',').Any(x=> arrSearchFilter.Contains(x))
This is not a good idea if the Product table is large, as you are losing a key benefit of SQL and loading ALL the data before filtering.
Redesign
If that is a problem and you can change the database structure, you should create a child table that replaces the comma-separated values with a proper normalised structure. Comma separated variables might look like a convenient shortcut but they are not a good design and as you have found, are not easy to work with in SQL.
SQL
If the design cannot be changed and the table is large, then your only other option is to hand-roll the SQL and execute this directly, but this would lose some of the benefits of having Linq.
I currently am using the following code
$select = $this->select()
->setIntegrityCheck(false)
->from(array('st' => $this->_name))
->join(array('sp' => 'staff_permissions'), 'sp.staff_id = st.id and sp.pool_id = ' . $pool_id )
->join(array('p' => 'permissions'), 'p.id = sp.permission_id')
->where('staff_id = ?', $staff_id);
return $this->fetchAll($select)->toArray();
It combines three tables and returns the result. The 'st' table corresponds to one staff (so one row), and the other two tables correspond to multiple rows. So what I was hoping was to get a single object back such that the other two tables are arrays inside the object.
So as an example, I get back $row, so that $row->first_name is the name, but $row->permission_id is an array with all the ids in it.
Can that be done using the JOIN clause?
This query should be done in the lowest layer in your application, the next layer up the stack will be your mappers. In your mapper layer you can map to your Entities, which will be a 'staff' entity (object) which contains a collection for 'staff_permissions' & a collection for 'permissions'
model diagram:
-----------
| service | // business logic
-----------
|
-----------
| mapper | // maps external data to internal entities (or vice versa)
-----------
|
----------- ----------------------
|dao (sql)| -> | zend table gateway |
----------- ----------------------
mapper example code:
$staffEntity = new StaffEntity();
$staffEntity->setName($response['name']);
foreach($response['staff_permissions] as $permission) {
$permission = new Permission();
$permission->setName($permission['name']);
$permission->setRule($permission['rule']);
// ... etc ...
$staffEntity->addPermission($permission);
}
// ... same for permissions ...
return $staffEntity;