I'm implementing a pseudo hash-tagging system for the company I work at with their customer's facebook photos. A customer can upload a photo to the page, and a page admin can tag it with the hashtag for a product.
What I am trying to do is get all photos from the page that have a certain tag in the comments (for instance, get all photos from the company page with a comment containing only '#bluepants').
I am trying to make sure the Facebook API handles the heavy lifting (we'll cache the results), so I'd like to use FQL or the Graphs API, but I can't seem to get it working (my SQL is quite rusty after relying on an ORM for so long). I would prefer if it outputs as many results as possible, but I'm not sure if FB lets you do more than 25 at once.
This is going to be implemented in a sinatra site (I am currently playing around with the Koala gem, so bonus points if I can query using it)
Could anyone give me some guidance?
Thanks!
I've got something like this working in FQL/PHP. Here is my multiquery.
{'activity':
"SELECT post_id, created_time FROM stream WHERE source_id = PAGE_ID AND
attachment.fb_object_type = 'photo' AND created_time > 1338834720
AND comments.count > 0 LIMIT 0, 500",
'commented':
"SELECT post_id, text, fromid FROM comment WHERE post_id IN
(SELECT post_id FROM #activity) AND AND (strpos(upper(text), '#HASHTAG') >= 0",
'accepted':
"SELECT post_id, actor_id, message, attachment, place, created_time, likes
FROM stream WHERE post_id IN (SELECT post_id FROM #commented)
ORDER BY likes.count DESC",
'images':
"SELECT pid, src, src_big, src_small, src_width, src_height FROM photo
WHERE pid IN (SELECT attachment.media.photo.pid FROM #accepted)",
'users':
"SELECT name, uid, current_location, locale FROM user WHERE uid IN
(SELECT actor_id FROM #accepted)",
'pages':
"SELECT name, page_id FROM page WHERE page_id IN (SELECT actor_id FROM #accepted)",
'places':
"SELECT name, page_id, description, display_subtext, latitude, longitude
FROM place WHERE page_id IN (SELECT place FROM #accepted)"
}
To break this down:
#activity gets all stream objects created after the start date of my campaign that are photos and have a non-zero comment count. Using a LIMIT of 500 seems to return the maximum number of posts. Higher or lower values return fewer.
#commented finds the posts that have #HASHTAG in the text of one of their comments. Note, I'm not looking for a #, which is a reserved character in FQL. Using it may cause you problems.
#accepted gets the full details of the posts found in #commented.
#images gets all the details of the images in those posts. I have it on my todos to refactor this to use object_id instead of pid and try using the new real_width specification to make my layout easier.
#users and #pages get the details of the actor who originally posted the item. I now know I could have used the profile table to get this in one query.
#places gets the location details for geo-tagged posts.
You can see this in action here: http://getwellgabby.org/show-us-a-sign
Related
I have created a SQFlite database for my app which is working very well. Basically, the app is to create a list of my client's company details. The company name appears in an Expansion Tile, and once you open the tile, you see the rest of the company's details. I know how to arrange the tiles in either descending or ascending order, but is there a way to arrange according to alphabetic order based on the company name? Thus, can one arrange the tables in SQFlite in alphabetic order instead of ASC or DESC? When my client's details are added dynamically, I want them to be arranged in alphabetic order so that it makes more sense. Thank you so much for any help.. I have tried to find a comment on this on Stackoverflow, but don't seem to be able to find one.
Ok, I managed to figure this out. My question really shows my inexperience. Anyway, Query has a orderBy property. In my database I created the following code:
Future<void> _createDB(Database db, int version) async {
await db.execute('''
CREATE TABLE todo(
id INTEGER PRIMARY KEY AUTOINCREMENT,
title TEXT,
name TEXT,
phone TEXT,
fax TEXT,
email TEXT,
street TEXT,
city TEXT,
town TEXT,
code TEXT,
isExpanded INTEGER
orderBy used to be 'id DESC'. It should simply be changed to 'title ASC'. Thus, it will order the sequence by the title in alphabetic order from top to bottom. Here is the code:
Future<List<Todo>> getTodo() async {
final db = await database;
// query the database and save as list of maps
List<Map<String, dynamic>> items = await db.query(
'todo',
orderBy: 'title ASC',
);
Thus, the correct question is not how to order in alphabetic order, but how to order based on title (which is the company name in my example) in ASCending order.
In my website I track clicks on a specific link and save the count on database with this function:
public function redirect($url)
{
$data = Class::where('url', $url)->first('clicks');
$cnt = $data->clicks;
$cnt++;
$sql = Class::where('url', $url)->update([
'clicks' => $cnt,
]);
if (!$sql) {
abort(500);
}
return redirect('https://website.com/'.$url);
}
the problem is if someone spam the link , the count goes up. I want stop count if user clicked on link 5mins ago.
he will redirected to the link but count doesn't go up.
I'm new so it's so good if you explain it with details. Thanks
I would create a new table, lets call it link_clicks. You will need 3 columns, one to identify the person, one to get the time and one to identify the link (I dont exactly know how you store the links you want to observe).
So more or less you will have the following table:
link_clicks
user_id
link_id
created_at
If the users are always logged in, I would store the user_id, if they are not logged in, I would store the IP-address. So instead of user_id make a column called ip or ip-address.
Afterwards you can easily get the last click and its time.
--Example
Your new table contains following entry:
user_id = 1, link_id = 1 and created_at = 2021-04-21 19:00:00
Now in your controller you get the current date_time date("Y-m-d H:i:s") and the user id like that: auth()->id(). You can also define your time treshold in minutes e.g max_dif = 5.
Afterwards you can query the table for all clicks for the user. You can either make the time comparision in your where() statement or you can make some comaprison in your php code to check if the time treshhold is reached.
I am designing a system that handles a large number of buried point event. An event record contains:
buried_point_id, for example: 1 means app_launch, 2 means user_register.
happened_at: the event timestamp.
user_id: the user identifier.
other attributes, including basic ones (phone_number, city, country) and user-defined ones (click_item_id, it literally can be any context information). PMs will add more and more user-defined attributes to the event record.
The query pattern is like:
SELECT COUNT(DISTINCT user_id) FROM buried_points WHERE buried_point_id = 1 AND city = 'San Francisco' AND click_item_id = 123;
Since my team invests heavily in ClickHouse, I want to leverage ClickHouse for the problem. I wonder if it is a good practice to use the experimental Map data type to store all attributes in a MAP-type column such as {city: San Francisco, click_item_id: 123, ...}, or any other recommendation? Thanks.
I'm trying to build a query that shows only non-unique duplicates. I've already built a query that shows all the records coming into consideration:
SELECT tbl_tm.title, lp_index.starttime, musicsound.archnr
FROM tbl_tm
INNER JOIN musicsound on tbl_tm.fk_tbl_tm_musicsound = musicsound.pk_musicsound
INNER JOIN lp_index ON musicsound.pk_musicsound = lp_index.fk_index_musicsound
INNER JOIN plan ON lp_index.fk_index_plan = plan.pk_plan
WHERE tbl_tm.FK_tbl_tm_title_type_music = '22' AND plan.airdate
BETWEEN to_date ('15-01-13') AND to_date('17-01-13')
GROUP BY tbl_tm.title, lp_index.starttime, musicsound.archnr
HAVING COUNT (tbl_tm.title) > 0;
The corresponding result set looks like this:
title starttime archnrr
============================================
Pumped up kicks 05:05:37 0616866
People Help The People 05:09:13 0620176
I can't dance 05:12:43 0600109
Locked Out Of Heaven 05:36:08 0620101
China in your hand 05:41:33 0600053
Locked Out Of Heaven 08:52:50 0620101
It gives me music titles played between a certain timespan along with their starting time and archive ID.
What I want to achieve is something like this:
title starttime archnr
============================================
Locked Out Of Heaven 05:36:08 0620101
Locked Out Of Heaven 08:52:50 0620101
There would only be two columns left: both share the same title and archive number but differ in the time part. Increasing the 'HAVING COUNT' value will give me a zero-row
result set, since there aren't any entries that are exactly the same.
What I've found out so far is that the solution for this problem will most likely have a nested subquery, but I can't seem to get it done. Any help on this would be greatly appreciated.
Note: I'm on a Oracle 11g-server. My user has read-only privileges. I use SQL Developer on my workstation.
You can try something like this:
SELECT title, starttime, archnr
FROM (
SELECT title, starttime, archnr, count(*) over (partition by title) cnt
FROM (your_query))
WHERE cnt > 1
Here is a sqlfiddle demo
I have a database with customers orders.
I want to use Linq (to EF) to query the db to bring back the last(most recent) 3,4...n orders for every customer.
Note:
Customer 1 may have just made 12 orders in the last hr; but customer 2 may not have made any since last week.
I cant for the life of me work out how to write query in linq (lambda expressions) to get the data set back.
Any good ideas?
Edit:
Customers and orders is a simplification. The table I am querying is actually a record of outbound messages to various web services. It just seemed easer to describe as customers and orders. The relationship is the same.
I am building a task that checks the last n messages for each web service to see if there were any failures. We are wanting a semi real time Health status of the webservices.
#CoreySunwold
My table Looks a bit like this:
MessageID, WebserviceID, SentTime, Status, Message, Error,
Or from a customer/order context if it makes it easer:
OrderID, CustomerID, StatusChangedDate, Status, WidgetName, Comments
Edit 2:
I eventually worked out something
(Hat tip to #StephenChung who basically came up with the exact same, but in classic linq)
var q = myTable.Where(d => d.EndTime > DateTime.Now.AddDays(-1))
.GroupBy(g => g.ConfigID)
.Select(g =>new
{
ConfigID = g.Key,
Data = g.OrderByDescending(d => d.EndTime)
.Take(3).Select(s => new
{
s.Status,
s.SentTime
})
}).ToList();
It does take a while to execute. So I am not sure if this is the most efficient expression.
This should give the last 3 orders of each customer (if having orders at all):
from o in db.Orders
group o by o.CustomerID into g
select new {
CustomerID=g.Key,
LastOrders=g.OrderByDescending(o => o.TimeEntered).Take(3).ToList()
}
However, I suspect this will force the database to return the entire Orders table before picking out the last 3 for each customer. Check the SQL generated.
If you need to optimize, you'll have to manually construct a SQL to only return up to the last 3, then make it into a view.
You can use SelectMany for this purpose:
customers.SelectMany(x=>x.orders.OrderByDescending(y=>y.Date).Take(n)).ToList();
How about this? I know it'll work with regular collections but don't know about EF.
yourCollection.OrderByDescending(item=>item.Date).Take(n);
var ordersByCustomer =
db.Customers.Select(c=>c.Orders.OrderByDescending(o=>o.OrderID).Take(n));
This will return the orders grouped by customer.
var orders = orders.Where(x => x.CustomerID == 1).OrderByDescending(x=>x.Date).Take(4);
This will take last 4 orders. Specific query depends on your table / entity structure.
Btw: You can take x as a order. So you can read it like: Get orders where order.CustomerID is equal to 1, OrderThem by order.Date and take first 4 'rows'.
Somebody might correct me here, but i think doing this is linq with a single query is probably very difficult if not impossible. I would use a store procedure and something like this
select
*
,RANK() OVER (PARTITION BY c.id ORDER BY o.order_time DESC) AS 'RANK'
from
customers c
inner join
order o
on
o.cust_id = c.id
where
RANK < 10 -- this is "n"
I've not used this syntax for a while so it might not be quite right, but if i understand the question then i think this is the best approach.