I want to use distinct function in users collection. In mongo shell, I
can use like the following:
db.users.distinct("name");
where name is the collections field used to distinct.
likewise I want, the same in c code. Only the distinct part I want,
familiar with creating connections, working cursors, working bson
data. Also, the link having only the
basics of mongo c API.
I need the complete documentation of c api, if anyone knows give me
the link.
thanks,
Seems like you can do that using the mongo_run_command function, since distinct is a command: db.runCommand( { distinct: 'users', key: 'name' } )
Related
I'm trying to get a list of domain names from email addresses using the Salesforce query language. This potentially really simple and something I would normally accomplish with split_part in postgresql, like:
SELECT split_part(Email, '#', 2)
FROM Lead
GROUP BY 1
I've been digging through the SOQL documentation and can't really find any standard string functions. However, there's this salesforce community answer: https://success.salesforce.com/answers?id=90630000000gi8EAAQ which uses LEFT, FIND and SUBSTITUTE. I tried something as simple as:
SELECT LEFT(Email, 3) FROM Lead limit 10
But get:
Error: MALFORMED_QUERY
Message: SELECT LEFT(Email, 3) FROM Lead limit 10
ERROR at Row:1:Column:17\nunexpected token: ','
Have these functions been deprecated?
I have lots of potential domain names and don't really want to query for pages and pages of every possible email address or I'll quickly hit my Salesforce API limit.
Those functions are not available in SOQL/SOSL. The link you provided refers to Custom Formula Fields. Are you trying to get the data using the UI or the API? Without knowing how you are extracting the data, the following are general suggestions:
You can create a formula field on your object called Email_Domain__c. Then, you can query or filter by that field.
You can also use custom formulas in Reports to filter your results.
Use Apex and/or Visualforce to extract/display/export the data.
Use the LIKE keyword to filter by domain name as in:
SELECT Email FROM LEAD WHERE Email LIKE '%gmail.com%'
The LIKE keyword is not efficient. https://trailhead.salesforce.com/en/modules/database_basics_dotnet/units/writing_efficient_queries
I'm very new to CouchDB and I have a simple task that I have no been able to find a straight answer.
I have two related docs: order and invoice
invoice = {
id: "invoice_id",
order: "order_id",
values: [...]
}
order = {
id: "order_id",
**order_number: 12345**
}
I have defined a map function that select the unfulfilled invoices, now I need the order_number, which is in the order doc. They are the same transaction. How do I fetch the order_number from the order when I get my invoices?
I've looked around and I'm getting so many answers like: view collation, linked documents, include_docs=true, structure docs to have both...
I'm just looking for the simplest way with a clear explanation. I appreciate any help.
p.s.
Since I'm new I'm finding couchDB development to be very involved. I have map functions, but they need to be pushed to the couchInstance? Or I edit the map functions in Futon? Are there better ways to develop against couchDB? I see there's couchApp but the docs are sparse and the project hasn't been updated in a while.
I have a requirement where the records will be sorted based on created date first and if created dates are same, we will sort on another field called as ratings.
In my Spring mongo project I am doing the following thing:
Query query = new Query();
query.with(new Sort(Direction.DESC, "crDate")).with(new Sort(Direction.DESC, "ratings"));
For some reasons its only sorting on the first field ie crDate. And if both dates are same, sort by ratings never work.
When i try to check the value of sort object it shows me this:
{"crDate":-1,"ratings":-1}
Another finding is, mongo takes in the following syntax for compound sorts:
db.abc.find({..some criteria..}).sort([["crDate",-1],["ratings",-1]]);
Is this a bug in spring mongodb implementation or I missed something?
Looking at the Spring API Documentation it shows you can specify multiple strings to the sort object you are creating in a list. From you snippet above I would suggest you need to only apply the one sort object that takes the two fields, something like
query.with(new Sort(Direction.DESC, Arrays.asList("crDate", "ratings")));
There was another constructor that took the List of Order objects. Strange but I tried it with that now and it seems to be working.
I am now using a single with clause and passing in a List of Order
I have following query:
Article.joins(:themes => [:users]).where(["articles.user_id != ?", current_user.id]).order("Random()").limit(15).uniq
and gives me the error
PG::Error: ERROR: for SELECT DISTINCT, ORDER BY expressions must appear in select list
LINE 1: ...s"."user_id" WHERE (articles.user_id != 1) ORDER BY Random() L...
When I update the original query to
Article.joins(:themes => [:users]).where(["articles.user_id != ?", current_user.id]).order("Random()").limit(15)#.uniq
so the error is gone... In MySQL .uniq works, in PostgreSQL not. Exist any alternative?
As the error states for SELECT DISTINCT, ORDER BY expressions must appear in select list.
Therefore, you must explicitly select for the clause you are ordering by.
Here is an example, it is similar to your case but generalize a bit.
Article.select('articles.*, RANDOM()')
.joins(:users)
.where(:column => 'whatever')
.order('Random()')
.uniq
.limit(15)
So, explicitly include your ORDER BY clause (in this case RANDOM()) using .select(). As shown above, in order for your query to return the Article attributes, you must explicitly select them also.
I hope this helps; good luck
Just to enrich the thread with more examples, in case you have nested relations in the query, you can try with the following statement.
Person.find(params[:id]).cars.select('cars.*, lower(cars.name)').order("lower(cars.name) ASC")
In the given example, you're asking all the cars for a given person, ordered by model name (Audi, Ferrari, Porsche)
I don't think this is a better way, but may help to address this kind of situation thinking in objects and collections, instead of a relational (Database) way.
Thanks!
I assume that the .uniq method is translated to a DISTINCT clause on the SQL. PostgreSQL is picky (pickier than MySQL) -- all fields in the select list when using DISTINCT must be present in the ORDER_BY (and GROUP_BY) clauses.
It's a little unclear what you are attempting to do (a random ordering?). In addition to posting the full SQL sent, if you could explain your objective, that might be helpful in finding an alternative.
I just upgraded my 100% working and tested application from 3.1.1 to 3.2.7 and now have this same PG::Error.
I am using Cancan...
#users = User.accessible_by(current_ability).order('lname asc').uniq
Removing the .uniq solves the problem and it was not necessary anyway for this simple query.
Still looking through the change notes between 3.1.1 and 3.2.7 to see what caused this to break.
I use Objectify for datastore operations in my GAE/Java application. I have used Objectify's #Embeded facility in a couple of places in my project. Objectify automatically flattens the nested objects within the entity marked by #Embeded notation using the . separator. Thus I have ended up with column names like entity.embededObject.Field
For example I have an entity 'Person' in my data store with two columns name and address.email.
I want to filter through Person in the datastore viewer by writing a simple GQL query.
But the following query fails with a syntax error:
SELECT * FROM Person where address.email='mail#gmail.com'
whereas the following works as it should
SELECT * FROM Person where name='Joe'
What am I doing wrong?
GQL currently doesn't support this - only 'word' characters are supported. You should definitely file this as a bug in the issue tracker.
Tested today, it is possible to run the following with backquotes
SELECT * FROM `your.kind`
I believe this holds true for any parameter, but please correct me if I am wrong.