Closed. This question is opinion-based. It is not currently accepting answers.
Closed 8 years ago.
Locked. This question and its answers are locked because the question is off-topic but has historical significance. It is not currently accepting new answers or interactions.
I have a function, parseQuery, that parses a SQL query into an abstract representation of that query.
I'm about to write a function that takes an abstract representation of a query and returns a SQL query string.
What should I call the second function?
I think the verb you want is 'compose'.
The opposite of parse is serialize
In compiler terminology, the opposite is "unparse". Specifically, parsing turns a stream of tokens into abstract syntax trees, while unparsing turns abstract syntax trees into a stream of tokens.
Compose? When parsing a query you break it into its constituent parts (tokens, etc.), the reverse would be composing the parts into a string query.
To complement your existing naming, composeQuery looks best.
But in the general case, the opposite of parse is ǝsɹɐd
I would use one of these:
ToString()
ToSQL()
Render()
I think "serialize" is probably the word you want. It means to produce a textual representation of data that can be exported (and imported) from the program.
The antonym of 'analyze' is 'synthesize'.
ToQueryString()
Definitely Render.
I would call it constructQuery.
generate or emit, possibly.
Just to add some stuff.
Surely parse is a two way word.
You can parse an abstract into a query.
You can parse a query into an abstract.
The question should be, what do you name the latter part of the method, and because in this instance you're parsing an abstract to make a query you'd call it parseAbstract.
To answer the question, parsing has no opposite.
generateQuery, possibly? createQuery?
Take your pick
Generate
Dump
Serialize
Emit
They each have slightly different connotations.
Maybe prettyPrintQuery?
compose, construct, generate, render,condense, reduce, toSQL, toString depending on the nature of the class and its related operators
A traditional compiler has two parts: a parser and a code generator.
So you could call it "Generate". Of course, it's a little bit different here because the compiler isn't writing source code. (unless it's a precompiler).
Possibly Format(). or ToSQL() in your instance?
unParse()? Just kidding, I would go with toQueryString()
flatten?
The parsed query object perhaps represents a condition hierarchy, which you are "flattening" back into a 1 dimensional string.
But given that you're going from object to string, really just use toString or toSQL() or something like that. Besides, if you designed it well and are using the right app, you can rename it later and just stick stuff in the comments on what it does.
I'd say serialize and deserialize, instead of parse and ...
I would go for ToString(), since you can usually chain-nest them (opposite functions, that let you pass from Class1 to Class2 and vice-versa)
DateTime.Parse( DateTime.Parse( myDate.ToString() ).ToString() );
Serialize() looks like a nice choice, but it already has an opposite in Deserialize().
In your specific scenario, as other pointed out, ToSql() is another good choice.
I'd use render
> a = 'html': { 'head': {'title': 'My Page'}, 'body': { 'h1': 'Hello World', 'p': 'This is a Paragraph' } }
> b = render(a)
> console.log(b)
<html>
<head>
<title>My Page</title>
</head>
<body>
<h1>Hello World</h1>
<p>This is a Paragraph</p>
</body>
</html>
Which is IMHO, the opposite to parse()
> c = parse(b)
{ 'html': {
'head': {
'title': 'My Page'
}
'body': {
'h1': 'Hello World',
'p': 'This is a Paragraph'
}
}
+1 for Generate, but tack on what you're generating, i.e. GenerateSQL()
I voted for 'compose' but if you don't like that I would also suggest 'build'
What about asSQL() or even more asQuery()?
INHO Serialize, synthesize are good options. Also, as you have named parseQuery, i will go with codeQuery
I usually use "parse" as a conversion method and, therefore, i can't find a opposite word for "convert". (you can't "deconvert" something, as "unconvert" is a type of conversion itself).
thinking this way, the best solution (for me) is having two "parse" methods that receive different arguments. Example (Java):
public class FooBarParser{
public Foo parse(Bar bar);
public Bar parse(Foo foo);
}
deparse
Deparse is to parse, as:
decompile is to compile
decompose is to compose
deserialize is to serialize
degroovy is to groovy :) ;)
Parsing / deparsing is not change of structure, but conversion. Precise conversion between equivalent text and abstract-syntax-tree formats, maintaining all relationships & structure.
"Compose" means change of structure, so is not quite right. It suggests combining from separate independent parts (usually for the first time). Just as "decompose" suggests splitting into independent parts. They change form, not just format.
A quick search show's the term's used within:
Perl: http://perldoc.perl.org/B/Deparse.html
R: http://www.hep.by/gnu/r-patched/r-lang/R-lang_98.html
Common Lisp: http://www.clisp.org/impnotes/dffi.html#c-type-parse
PostgreSQL: http://doxygen.postgresql.org/deparse_8c.html
Eclipse: http://www.eclipse.org/forums/index.php/t/201883/
Unix Korn Shell: http://www.sourcecodebrowser.com/ksh/93tplus-p/deparse_8c.html
Related
Suppose I have this code:
(handler-case (read ...)
(parse-error (condition)
(format t "What text was I reading last to get this error? ~s~&"
(how-to-get-this-text? condition))))
I can only see the parse-namestring accessors, but it gives the message of the error, not the text it was parsing.
EDIT
In my case the problem is less generic, so an alternative solution not involving the entire string that failed to parse can be good too.
Imagine this example code I'm trying to parse:
prefix(perhaps (nested (symbolic)) expressions))suffix
In some cases I need to stop on "suffix" and in others, I need to continue, the suffix itself has no other meaning but just being an indicator of the action the parser should take next.
READ parses from a stream, not a string. The s-expression can be arbitrarily long. Should READ keep a string of what's been read?
What you might need is a special stream. In standard Common Lisp there is no mechanism for user defined streams. But in real life every implementation has such extensible streams. See for example 'gray streams'.
http://www.sbcl.org/1.0/manual/Gray-Streams.html
There's no standard function to do it. You might be able to brute-force something with read-from-string, but whatever you do, it will require some extra work.
Using ActiveRecord::Base.to_json I do:
user = User.find_by_name 'Mika'
{"created_at":"2011-07-10T11:30:49+03:00","id":5,"is_deleted":null,"name":"Mika"}
Now, what I would like to have is:
{
"created_at":"2011-07-10T11:30:49+03:00",
"id":5,
"is_deleted":null,
"name":"Mika"
}
Is there an option to do this?
It would be great to have a global option, so that the behaviour be set depending on dev/live environment.
I'll go out on a limb and say "no, there is no such option".
AFAIK, the JSON encoding is actually handled by ActiveSupport rather than ActiveRecord. If you look at lib/active_support/json/encoding.rb for your ActiveSupport gem, you'll see a lot of monkey patching going on to add as_json and encode_json methods to some core classes; the as_json methods are just used to flatten things like Time, Regexp, etc. to simpler types such as String. The interesting monkey patches are the encode_json ones, those methods are doing the real work of producing JSON and there's nothing in them for controlling the final output format; the Hash version, for example, is just this:
# comments removed for clarity
def encode_json(encoder)
"{#{map { |k,v| "#{encoder.encode(k.to_s)}:#{encoder.encode(v, false)}" } * ','}}"
end
The encoder just handles things like Unicode and quote escaping. As you can see, the encode_json is just mashing it all together in one compact string with no options for enabling prettiness.
All the complex classes appear to boil down to Hash or Array during JSONification so you could, in theory, add your own monkey patches to Hash and Array to make them produce something pretty. However, you might have some trouble keeping track of how deep in the structure you are so producing this:
{
"created_at":"2011-07-10T11:30:49+03:00",
"id":5,
"is_deleted":null,
"name":"Mika"
"nested":{
"not":"so pretty now",
"is":"it"
}
}
Would be pretty straight forward but this:
{
"created_at":"2011-07-10T11:30:49+03:00",
"id":5,
"is_deleted":null,
"name":"Mika"
"nested": {
"not":"so pretty now",
"is":"it"
}
}
would be harder and, presumably, you'd want the latter and especially so with deeply nested JSON where eyeballing the structure is difficult. You might be able to hang a bit of state on the encoder that gets passed around but that would be getting a little ugly and brittle.
A more feasible option would be an output filter to parse and reformat the JSON before sending it off to the browser. You'd have to borrow or build the pretty printer but that shouldn't be that difficult. You should be able to conditionally attach said filter only for your development environment without too much ugliness as well.
If you're just looking to debug your JSON based interactions then maybe the JSONovich extension for Firefox would be less hassle. JSONovich has a few nice features (such as expanding and collapsing nested structures) that go beyond simple pretty printing too.
BTW, I reviewed Rails 3.0 and 3.1 for this, you can check Rails 2 if you're interested.
Have you considered the JSON gem? I believe it does exactly what you're looking for.
e.g.
JSON.pretty_generate(user)
Have a look at the detail here...
http://apidock.com/ruby/JSON/pretty_generate
I think there must be a name for whatever it is I'm trying to do. I want to have a list of... requests and what the responses should look like. But the different requests and responses are not formatted the same. They are simple strings, so here's some examples:
request => response
foo%% => "OK"
foo? => "%%"
%%%%bar => "OK"
Version? => "Baz-%%%%"
Where % could be a number or letter.
I can code for each possible command with a big switch/case, but I wanted to make it more extend-able and maybe testable. Sorry this question is so vague. Please rename it and/or tag it correctly.
I'm doing this in javascript/node if it matters.
I don't know what you're inventing but it sounds like regular expressions.
You would call % a wildcard in this case.
Your code will need parse the input to verify syntactic correctness and output a data structure that represents the input. Then you will need to analyze your intermediate data structure to determine semantics. The simplest way would be to use objects that implement a common interface.
object FooRequest (value)
implements Request
method Process ()
return "OK"
I am working on a Symfony project and I currently have this:
<?php echo preg_replace('/\n/','<br />', $review->getComments()); ?>
and would very much like to be able to make all getters add html line breaks so i don't have to pepper my code with preg_replace. the $object->getFieldname methods are work automatically so I am looking to extend this somewhere to globally add a new method. What is the best approach here?
Seems like everyone forgot about nl2br() which is a function that does exactly that in PHP.
nl2br($review->getComments());
EDIT: At the time of this writing, everyone else uses preg_replace().
I think the best idea would be to add a getCommentsHtml() method onto your review object, which does something like:
return preg_replace('/\n/','<br />', $this->getComments());
Then you can use $review->getCommentsHtml() to format them using HTML. Also as Charlie mentioned, maybe str_replace would be better to use, as using a regular expression to change \n's into <br />'s may be a little bit of overkill :)
So if you don't want to have your code littered with replaces like this, I think putting a helper method on the classes that you'd like to format nicely would be the best way to go :)
How about:
str_replace("\n",'<br />', $review->getComments());
does anyone knows a file format for configuration files easy to read by humans? I want to have something like tag = value where value may be:
String
Number(int or float)
Boolean(true/false)
Array(of String values, Number values, Boolean values)
Another structure(it will be more clear what I mean in the fallowing example)
Now I use something like this:
IntTag=1
FloatTag=1.1
StringTag="a string"
BoolTag=true
ArrayTag1=[1 2 3]
ArrayTag2=[1.1 2.1 3.1]
ArrayTag3=["str1" "str2" "str3"]
StructTag=
{
NestedTag1=1
NestedTag2="str1"
}
and so on.
Parsing is easy but for large files I find it hard to read/edit in text editors. I don't like xml for the same reason, it's hard to read. INI does not support nesting and I want to be able to nest tags. I also don't want a complicated format because I will use limited kind of values as I mentioned above.
Thanks for any help.
What about YAML ? It's easy to parse, nicely structured has wide programming language support. If you don't need the full feature set, you could also use JSON.
Try YAML - is (subjectively) easy to read, allows nesting, and is relatively simple to parse.