How to do "Compile time" initialization of record array?

  • Thread starter Thread starter John Kelsey
  • Start date Start date
J

John Kelsey

Back in the "old" days with C, I used to do something like...

struct
{
char Description[20];
float price;
} Items[] =
{
{"Apple", 1.99},
{"Banana", 2.04}
};

If I wanted to do something similiar (and I do), does anyone have hints for
an obviously new C# programmer?

Any help is greatly appreciated,
John Kelsey
 
That looks like a hashtable. So i would do...

Hashtable hashtable = new Hashtable();
hashtable.add("Apple", 1.99);
hashtable.add("Banana", 2.04);

However you will have to know the key (e.g. 'Apple' or 'Banana') to
refer to these items in the hashtable. I would recomend creating a new
object with properties that you desire. Then create an ArrayList and
add that object to the ArrayList.
 
Dave,

I appreciate the response but I was looking for something slightly
different.

My array was initialized before runtime. I understand that I can add
records to a "collection" as part of the program execution, but in C, the
array contained the values I wanted without having to write one executable
statement.

This may be a subtle point that C# doesn't support... it may just be my
programming style...

I guess it boils down to the difference between
int i = 32;

and
int i;
....
i = 32;


Is this any clearer?
 
John Kelsey said:
Back in the "old" days with C, I used to do something like...

struct
{
char Description[20];
float price;
} Items[] =
{
{"Apple", 1.99},
{"Banana", 2.04}
};

If I wanted to do something similiar (and I do), does anyone have hints for
an obviously new C# programmer?
Click to expand...

Well, firstly I'd separate the declaration of the type from the use of
it. I'd then provide a constructor which took the description and the
price, and use that. For instance:

using System;

public class Item
{
string description;
public string Description
{
get { return description; }
}

// Note change to use the decimal type,
// which is usually more appropriate than float
// for money
decimal price;
public decimal Price
{
get { return price; }
}

public Item (string description, decimal price)
{
this.description = description;
this.price = price;
}
}

public class Test
{
static void Main()
{
Item[] items =
{
new Item ("Apple", 1.99m),
new Item ("Banana", 2.0m)
};

foreach (Item item in items)
{
Console.WriteLine ("{0} costs {1}",
item.Description,
item.Price);
}
}
}
 
if you want it to be an array of Veggie Structs, something like:

struct Veggie
{
string Description;
decimal Price;
public Veggie(string description, decimal price)
{
Description =description;
Price = price;
}
}

//usage:
Veggie[] Items =new Veggie[] {new Veggie("Pear",2.99), new Veggie("Avocado",
5.13), ....};

(untested)
--Peter
 
John Kelsey said:
I appreciate the response but I was looking for something slightly
different.

My array was initialized before runtime. I understand that I can add
records to a "collection" as part of the program execution, but in C, the
array contained the values I wanted without having to write one executable
statement.

This may be a subtle point that C# doesn't support... it may just be my
programming style...

I guess it boils down to the difference between
int i = 32;

and
int i;
...
i = 32;
Click to expand...

Well, in that case there really isn't any difference in the compiled
code, as far as I'm aware. I know what you mean though :)
Is this any clearer?
Click to expand...

While I see what you mean, there isn't a way of doing it that I know of
in C#, nor do I suspect it's really a problem. Do you have any
particular reason for desiring it from an execution point of view? You
can achieve the same "end goal" (getting an initialised array) in C#
with quite similar code, so it's only the matter of when it actually
gets built which is significantly different. Are you worried about
performance?
 
John,

If I'm understanding you correctly, what you want to do is have an
array established before your program runs. Is this correct? I think
I was trying to do something like this a few days ago. I wanted to
have a constant hashtable. Since I don't think this is possible, I
created an object that was going to contain this hashtable. Then I
declared all the constants that I wanted to be in the hashtable and
then in the contructor I added them to the hashtable. Then when ever I
call this object the hashtable is already filled. You can call
initialize this object at load time so that the array is cached for the
remaining run time. I made this object a singleton because everything
in this object is constant. This keeps my cache smaller for better
performance. Hope this helps.
 
Jon,

For example, one of my "C" programs has

struct
{
double L;
double M;
double S;
} LengthArray[] = {
{ 1.267004226, 49.988884080, 0.053112191}, // 1
{ 0.511237696, 52.695975300, 0.048692684}, // 2
{ -0.452244460, 56.628428550, 0.044116830}, // 3

This is some statistical data (there are 76 records per array) and I have
Length, Weight, and Circumference arrays.

I typically am targeting some low "horsepower" handheld devices.

Even in C, I could run time initialize my arrays but I didn't for at least a
couple of reasons...
1) Performance, why waste runtime cycles building an array when it could be
pre-built by the compiler
2) Run time building meant that I carried two copies of my data in the
"execution footprint". One copy in code segment and the other copy in the
data segment.

I understand that this really doesn't present a challenge to a modern
device, especially in a desktop environment, it was just a style I have/had.

I don't understand C# the way I do C. Things that used to be important, no
longer seem to matter as much.

I've appreciated all of the responses and I'm certain I'll ask more "old
fashioned" questions as I learn.

Thanks,
John Kelsey
 
John Kelsey said:
For example, one of my "C" programs has

struct
{
double L;
double M;
double S;
} LengthArray[] = {
{ 1.267004226, 49.988884080, 0.053112191}, // 1
{ 0.511237696, 52.695975300, 0.048692684}, // 2
{ -0.452244460, 56.628428550, 0.044116830}, // 3

This is some statistical data (there are 76 records per array) and I have
Length, Weight, and Circumference arrays.

I typically am targeting some low "horsepower" handheld devices.

Even in C, I could run time initialize my arrays but I didn't for at least a
couple of reasons...
1) Performance, why waste runtime cycles building an array when it could be
pre-built by the compiler
2) Run time building meant that I carried two copies of my data in the
"execution footprint". One copy in code segment and the other copy in the
data segment.
Click to expand...

In both of these cases, I'd expect the time and memory taken to be
really pretty tiny. You'll be wasting a few K and a couple of
milliseconds (once), even on a handheld device. These days, that's not
likely to be a significant problem.
I understand that this really doesn't present a challenge to a modern
device, especially in a desktop environment, it was just a style I have/had.
Sure.

I don't understand C# the way I do C. Things that used to be important, no
longer seem to matter as much.
Click to expand...

Indeed. However, the things which were a real pain in C are much easier
in C# and .NET. Modern languages tend to value the time of the
developer over the time of the processor, on the grounds that it's
generally a lot more expensive :)
I've appreciated all of the responses and I'm certain I'll ask more "old
fashioned" questions as I learn.
Click to expand...

Goodo. Do keep 'em coming.
 
I never learned C. I started programming with the introduction of Java
so I'm sure im unfamiliar with old techniques. But I dont think there
is a way to initialize an array at compile time in C#. I have a
question about C since i am unfamiliar to it. How does it fill an
array at compile time, isn't an array just a cached set of objects? If
it caches the objects at compile time then that means only the machine
that compiles the program can access the array and only after it has
been compiled.

About your program. I think your best bet is to create and object with
different properties and then add them to an ArrayList. e.g.....

public class Veggies
{
private static int L;
private static int M;
private static int S;
public Veggies(int l, int m, int s)
{
L = l;
M = m;
S = s;
}
public static int L()
{
get
{return L;}
}
...
...
}

//in your other class
private ArrayList list = new ArrayList();
Veggies oVeggies = new Veggies(1.267004226, 49.988884080, 0.053112191);
list.Add(oVeggies);
Veggies oVeggies = new Veggies(0.511237696, 52.695975300, 0.048692684);
list.Add(oVegges);
//...
//and so on...

Hope this helped.
 
I believe both Pears and Avocados are technically fruit :)

Peter Bromberg said:
if you want it to be an array of Veggie Structs, something like:

struct Veggie
{
string Description;
decimal Price;
public Veggie(string description, decimal price)
{
Description =description;
Price = price;
}
}

//usage:
Veggie[] Items =new Veggie[] {new Veggie("Pear",2.99), new
Veggie("Avocado",
5.13), ....};

(untested)
--Peter
--
Co-founder, Eggheadcafe.com developer portal:
http://www.eggheadcafe.com
UnBlog:
http://petesbloggerama.blogspot.com




John Kelsey said:
Back in the "old" days with C, I used to do something like...

struct
{
char Description[20];
float price;
} Items[] =
{
{"Apple", 1.99},
{"Banana", 2.04}
};

If I wanted to do something similiar (and I do), does anyone have hints
for
an obviously new C# programmer?

Any help is greatly appreciated,
John Kelsey
Click to expand...
Click to expand...
 
Dave,

The answer to you question is a little complicated and probably out of scope
for this newsgroup, but...
You could simplistically think of old C programs as having:
1) a Data segment, "long term" variable storage but only lasts as long as
the program (think global variables)
2) a Code segment, executable statements (mostly code, but some data)
3) a stack for "short term" variable storage

For compiler initialized variables, the compiler would create an image of
the data segment that had certain values already set. The program loader
would load the program with the initialized data segment and the first
executable statement in the program had access to the data segment
image/compiler initialized variables.

All of the work is done by the compiler and the program loader, not by the
code I write.

This is highly simplified, but I hope you get the point. A point of
interest but probably only of value if you ever take ancient computer
languages subject.

Thanks for the help.
John Kelsey

Dave said:
I never learned C. I started programming with the introduction of Java
so I'm sure im unfamiliar with old techniques. But I dont think there
is a way to initialize an array at compile time in C#. I have a
question about C since i am unfamiliar to it. How does it fill an
array at compile time, isn't an array just a cached set of objects? If
it caches the objects at compile time then that means only the machine
that compiles the program can access the array and only after it has
been compiled.
Click to expand...
--snip--
 
You don't have to go to Arizona to find Yuma.
--
Co-founder, Eggheadcafe.com developer portal:
http://www.eggheadcafe.com
UnBlog:
http://petesbloggerama.blogspot.com




Marina said:
I believe both Pears and Avocados are technically fruit :)

Peter Bromberg said:
if you want it to be an array of Veggie Structs, something like:

struct Veggie
{
string Description;
decimal Price;
public Veggie(string description, decimal price)
{
Description =description;
Price = price;
}
}

//usage:
Veggie[] Items =new Veggie[] {new Veggie("Pear",2.99), new
Veggie("Avocado",
5.13), ....};

(untested)
--Peter
--
Co-founder, Eggheadcafe.com developer portal:
http://www.eggheadcafe.com
UnBlog:
http://petesbloggerama.blogspot.com




John Kelsey said:
Back in the "old" days with C, I used to do something like...

struct
{
char Description[20];
float price;
} Items[] =
{
{"Apple", 1.99},
{"Banana", 2.04}
};

If I wanted to do something similiar (and I do), does anyone have hints
for
an obviously new C# programmer?

Any help is greatly appreciated,
John Kelsey
Click to expand...
Click to expand...
Click to expand...
 
It should be noted that there is proposed syntax for C# 3.0 which will
allow initialization with the following syntax:

public struct MyStruct
{
public double L;
public double M;
public double S;
}

MyStruct[] lengthArray[] = new MyStruct[]
{
{ L = 1.267004226, M = 49.988884080, S = 0.053112191},
{ L = 0.511237696, M = 52.695975300, S = 0.048692684},
{ L = -0.452244460, M = 56.628428550, S = 0.044116830}
}

I might have not gotten it completely correct, but the syntax is
something like that.

What actually gets compiled is the same array initialization as there is
now, but code that will actually populate the properties/fields like regular
assignments. It's been a while since I used the C# 3.0 preview, but the
proposed spec amendments go into much more detail.


--
- Nicholas Paldino [.NET/C# MVP]
- (e-mail address removed)

Jon Skeet said:
John Kelsey said:
For example, one of my "C" programs has

struct
{
double L;
double M;
double S;
} LengthArray[] = {
{ 1.267004226, 49.988884080, 0.053112191}, // 1
{ 0.511237696, 52.695975300, 0.048692684}, // 2
{ -0.452244460, 56.628428550, 0.044116830}, // 3

This is some statistical data (there are 76 records per array) and I have
Length, Weight, and Circumference arrays.

I typically am targeting some low "horsepower" handheld devices.

Even in C, I could run time initialize my arrays but I didn't for at
least a
couple of reasons...
1) Performance, why waste runtime cycles building an array when it could
be
pre-built by the compiler
2) Run time building meant that I carried two copies of my data in the
"execution footprint". One copy in code segment and the other copy in
the
data segment.
Click to expand...

In both of these cases, I'd expect the time and memory taken to be
really pretty tiny. You'll be wasting a few K and a couple of
milliseconds (once), even on a handheld device. These days, that's not
likely to be a significant problem.
I understand that this really doesn't present a challenge to a modern
device, especially in a desktop environment, it was just a style I
have/had.
Sure.

I don't understand C# the way I do C. Things that used to be important,
no
longer seem to matter as much.
Click to expand...

Indeed. However, the things which were a real pain in C are much easier
in C# and .NET. Modern languages tend to value the time of the
developer over the time of the processor, on the grounds that it's
generally a lot more expensive :)
I've appreciated all of the responses and I'm certain I'll ask more "old
fashioned" questions as I learn.
Click to expand...

Goodo. Do keep 'em coming.
Click to expand...
 
John,

I've also learned C programming before I ever heard of .NET and I'm
always happy to help a fellow dinosaur. :)

One big difference between C and C# is that all .NET arrays are always
located on the (managed, garbage-collected) heap. There is no such
thing as a pre-allocated, pre-initialized array in C#.

You can have simple constants (including strings which are special in
..NET) that are evaluated at compile time and stored in the executable
image, but you can't have arrays that are stored in this way. (.NET
strings are _not_ character arrays, they are a separate data type.)

So while C# does syntactically allow initializing arrays in a single
big statement, you actually don't gain anything over assigning each
array element in a separate statement. The array itself is always
allocated at runtime on the heap, and all initialization statements
are therefore also evaluated at runtime. There's no way around this.

Hope this explains things,
Chris
 
int[] pins = new int[4] {9,3,7,2} ;

Damned if I know what really happens in the compiler, but it looks like the
above means the array is created at compile time but the address is
transferred to pins at execution. Or maybe not.
 
Back
Top