Squeel evaluating a string as a keypath for joins - ruby-on-rails

Here's what I'd really like to do:
Outage.joins{eval "unit.plant"}
I'm going to have a string that represents a squeel keypath. I'd like to send this string into a join statement.
But here's what I get as an error:
#<Class:0x639e328>: unknown class: Squeel::Nodes::Function
I've tried:
Outage.joins{"unit.plant"}
But that does not work... what am I missing?
I don't see any documentation on this on the github page: https://github.com/ernie/squeel/
Thank you for any help!
Update:
I've found a really confusing way to pull this off:
("unit.plant".split ".").map{ |x| x.to_sym }.reverse.inject{ |acc, x| { x => acc } }
This will output:
{:unit=>:plant}
This can be passed into the joins squeel function. I'm sure there has to be a better way than to construct nested symbols :(

How about
Outage.joins{Squeel::Nodes::KeyPath.new("unit.plant".split("."))}
Source: KeyPath Docs

Alternative solution, it can be chained with outer/inner:
strs = "unit.plant"
Outage.joins{strs.split(".").inject((strs.present? ? self : nil), :__send__)}
# chain with inner
# Outage.joins{strs.split(".").inject((strs.present? ? self : nil), :__send__).inner}

Related

Rails/Postgres nested JSON query

I have a JSON column in my events table called payload. I can create an event like this:
Event.create!(application_id: 1, stage: 'INITIATION', payload: { key: 'OUTCOME', value: {school_id: 2, prefered_language: 'GBP'}})
Simple queries like
Event.where("payload->>'key' = ?", "OUTCOME")
Will work fine, but how can I then add extra filters with the JSON that's nested in the value portion
Event.where("payload->>'key' = ? AND payload ->>'value'->>'school' = ?", "OUTCOME", 2)
^^^^^^^^^^^^^^^^^^^^
I tried that but i got:
PG::UndefinedFunction: ERROR: operator does not exist: text ->> unknown
LINE 1: ...ad ->>'key' = 'INPUTPARAMS' AND payload ->>'value'->>'school... ^
HINT: No operator matches the given name and argument type(s). You might need to add explicit type casts.
How do I get into the school attribute within value?
I have taken the idea from this documentation on ActiveRecord and JSON
After much searching, I came across this article.
The #>> operator allows you to dig inside the JSON. My query could be fixed like this.
Event.where("payload->>'key'= ? AND payload#>>'{value, school_id}' = ?", 'OUTCOME', shcool_id)
You could theoretically go down further by adding each layer to the predicate block {value, school, next_level, and_so_on}

Ruby, accessing a nested value in a hash

I have the following hash. Using ruby, I want to get the value of "runs". I can't figure out how to do it. If I do my_hash['entries'], I can dig down that far. If I take that value and dig down lower, I get this error:
no implicit conversion of String into Integer:
{"id"=>2582, "entries"=>[{"id"=>"7", "runs"=>[{"id"=>2588, ...
Assuming that you want to lookup values by id, Array#detect comes to the rescue:
h = {"id"=>2582, "entries"=>[{"id"=>"7", "runs"=>[{"id"=>2588}]}]}
# ⇓⇓⇓⇓⇓⇓⇓ lookup element with id = 7
h['entries'].detect { |e| e['id'] == 7 }['runs']
.detect { |e| e['id'] == 2588 }
#⇒ { "id" => 2588 }
As you have an array inside the entries so you can access it using an index like this:
my_hash["entries"][0]["runs"]
You need to follow the same for accessing values inside the runs as it is also an array.
Hope this helps.
I'm not sure about your hash, as it's incomplete. So , guessing you have multiple run values like:
hash = {"id"=>2582, "entries"=>[{"id"=>"7", "runs"=>[{"id"=>2588}]},
{"id"=>"8", "runs"=>[{"id"=>2589}]},
{"id"=>"9", "runs"=>[{"id"=>2590}]}]}
Then, you can do
hash["entries"].map{|entry| entry["runs"]}
OUTPUT
[[{"id"=>2588}], [{"id"=>2589}], [{"id"=>2590}]]

how to use ilike with Integer in grails

I use EasyGrid plugin and must find values where integer field like '%001%'
initialCriteria {
ilike('id', "%"+params.id+"%")
}
But ilike doesn't work with Integer. How to do it?
I tried to do:
initialCriteria {
ilike('id'.toString(), "%"+params.id+"%")
}
initialCriteria {
ilike('str(id)', "%"+params.id+"%")
}
but it's not work.
If id is an integer in the database, then ilike doesn't really make much sense and there is probably a better way to do what you are trying to do (like adding a type field or something to the domain object, and filter by type)
However, you should be able to do something like this (untested):
initialCriteria {
sqlRestriction "cast( id AS char( 256 ) ) like '%001%'"
}
Following criteria not working if you search from your textbox when user search any text character by mistake like 12dfdsf as your searchable id. It will give you an exception
initialCriteria {
ilike('id', "%"+params.id+"%")
}
For better use you can use following criteria
initialCriteria {
sqlRestriction "id like '%${params?.id}%'"
}
You could do:
String paddedId = params.id.toString().padLeft(3,'0')
initialCriteria {
ilike('id', "%$paddedId%")
}
The solution offered by tim_yates with the sqlRestriction would work in version 1.5.0 of easygrid.
One of the main differences from 1.4.x is that the gorm datasource no longer uses DetachedCriteria, but Criteria - which maps directly to Hibernate's Criteria API.
So you can try it on the last version.
(Keep in mind that the upgrade might break your existing grids. There's also many other changes)
Another small observation is that 'initialCriteria' is not the right place to do stuff like that. (it's not wrong, but there is a 'globalFilterClosure' property for applying column independent filters)
I mixed the code posted by #tim_yates and mine:
String paddedId = params.id.toString().padLeft(3,'0')
def crit = Book.withCriteria {
sqlRestriction "lpad(cast( id AS char( 256 ) ), 3, '0') like '%${paddedId}%'"
}
I've tried with a h2 in-memory db and it works, but I am not sure about two things:
the real usefulness of that
lpad syntax consistence across all db engines
YMMV

Mapping to the Keys of a Hash

I am working with a hash called my_hash :
{"2011-02-01 00:00:00+00"=>816, "2011-01-01 00:00:00+00"=>58, "2011-03-01 00:00:00+00"=>241}
First, I try to parse all the keys, in my_hash (which are times).
my_hash.keys.sort.each do |key|
parsed_keys << Date.parse(key).to_s
end
Which gives me this :
["2011-01-01", "2011-02-01", "2011-03-01"]
Then, I try to map parsed_keys back to the keys of my_hash :
Hash[my_hash.map {|k,v| [parsed_keys[k], v]}]
But that returns the following error :
TypeError: can't convert String into Integer
How can I map parsed_keys back to the keys of my_hash ?
My aim is to get rid of the "00:00:00+00" at end of all the keys.
Why don't you just do this?
my_hash.map{|k,v| {k.gsub(" 00:00:00+00","") => v}}.reduce(:merge)
This gives you
{"2011-02-01"=>816, "2011-01-01"=>58, "2011-03-01"=>241}
There is a new "Rails way" methods for this task :)
http://api.rubyonrails.org/classes/Hash.html#method-i-transform_keys
Using iblue answer, you could use a regexp to handle this situation, for example:
pattern = /00:00:00(\+00)+/
my_hash.map{|k,v| {k.gsub(pattern,"") => v}}.reduce(:merge)
You could improve the pattern to handle different situations.
Hope it helps.
Edit:
Sorry, iblue have already posted the answer
Another alternative could be:
map
return converted two elements array [converted_key, converted_value]
convert back to a hash
irb(main):001:0> {a: 1, b: 2}.map{|k,v| [k.to_s, v.to_s]}.to_h
=> {"a"=>"1", "b"=>"2"}

Why am I getting this TypeError - "can't convert Symbol into Integer"?

I have an array of hashes. Each entry looks like this:
- !map:Hashie::Mash
name: Connor H Peters
id: "506253404"
I'm trying to create a second array, which contains just the id values.
["506253404"]
This is how I'm doing it
second_array = first_array.map { |hash| hash[:id] }
But I'm getting this error
TypeError in PagesController#home
can't convert Symbol into Integer
If I try
second_array = first_array.map { |hash| hash["id"] }
I get
TypeError in PagesController#home
can't convert String into Integer
What am I doing wrong? Thanks for reading.
You're using Hashie, which isn't the same as Hash from ruby core. Looking at the Hashie github repo, it seems that you can access hash keys as methods:
first_array.map { |hash| hash.id }
Try this out and see if that works--make sure that it doesn't return the object_id. As such, you may want to double-check by doing first_array.map { |hash| hash.name } to see if you're really accessing the right data.
Then, provided it's correct, you can use a proc to get the id (but with a bit more brevity):
first_array.map(&:id)
This sounds like inside the map block that hash is not actually a hashie - it's an array for some reason.
The result is that the [] method is actually an array accessor method and requires an integer. Eg. hash[0] would be valid, but not hash["id"].
You could try:
first_array.flatten.map{|hash| hash.id}
which would ensure that if you do have any nested arrays that nesting is removed.
Or perhaps
first_array.map{|hash| hash.id if hash.respond_to?(:id)}
But either way you may end up with unexpected behaviour.

Resources