1. Java compile operation problem
2. Object operations easy to appear NullPointexception errors
3. Excess processing statement
4. Parameter delivery problem
5. Exception treatment
6. Database operation problem
7. INDEX Crossing
8. Others
9. Problem to be discussed
1.Java compilation problem
The Java program file first has to edit the Class file of the zodum code, and then run through the JVM. Java has a platform-independent because Sun provides a JVM (Java virtual machine) for most operating systems so as long as we use a unified API without care the underlying system.
In the early days of using Java, I am probably java.lang.classNotFoundException and java.lansNotFoundExceEction and java.lang.noclassdeffounderror, which is mainly the problem with ClassPath settings, similar to the dynamic link library inside C / C , if your Source Using other Package APIs so you have to set it to classpath when you have documented, you can point to a directory, generally to include the needs of the Classes, or point to a JAR or ZIP Documents, they are the package of classes packages. Such as:
Set classpath = c: /jdk1.3.1/lib/tools.jar; D: / wevelop / csc / class
Run under Windows Command
Set ClassPath
You can view the ClassPath currently set, if you want to add settings:
Set classpath =% classpath%; c: /bea/wlserver6.1/lib/weblogic.jar;
If you use the command line to make a documented, you have to set the JDK's PATH. Such as:
SET PATH = C: /JDK1.3.1/bin;% PATH%
Of course, if you are not too trouble, you can specify a full path:
C: /jdk1.3.1/bin/javac Yourown.java
C: /jdk1.3.1/bin/java Yourown
ClassPath can also be specified when completing. Such as:
Javac-ClassPath% my_classpath% Yourown.java
Java -classpath% my_classpath% Yourown
In general, when the JVM is running, there is a default LOAD Classes, you can run Java -verbose to view, generally% jdk_home% / jre / lib, I18n.jar below, rt.jar, etc. If you put your JAR file To the ext directory below this directory, you don't have to specify it, JVM automatically loads these JAR.
In addition, it is important to note that when running the class, it is a full name of the class, which contains its package name, if there is a class declaration as follows:
Package cn.com.sunjapan.util
Public class stringutil {
Public static void main (String [] args) {
System.out.println ("Hello World");
}
}
You have to use it when you run.
Java cn.com.sunjapan.util.stringutil
And you can't use Java Stringutil or you will have java.lang.noclassdefounderror.
2. Object Operation Easy NullPointexception Errors This error is probably the most probable error in the initial stage of programming. Java is an object-oriented language that is almost all between objects, and an instance of a class cannot call the method of this instance if it is a null. Otherwise, a java.lang.nullpointException error. The most common String operation, such as:
String str = NULL;
IF (Str.Equals ("Hello")) {
System.out.println ("Str Is Hello");
}
Commonly used to avoid it is to judge whether it is NULL before using an Object, unless you make sure it is definitely not null. The above example is modified as follows:
String str = NULL;
IF (str! = null && str.equals ("hello")) {
System.out.println ("Str Is Hello");
}
The operation of this equals or equalsignoreCase for String is often available in the following methods:
String str = NULL;
IF ("Hello". Equals (STR)) {
System.out.println ("Str Is Hello");
}
Use a determined String that is not null to compare with unknown String.
3. Excess processing statement
This problem is of course not just Java, any program may have excess garbage, although its final result is correct, we should avoid this unnecessary process as much as possible when writing procedures.
The common situation has the following:
More than 3.1 instance construction
Declare an instance of an object, some people like to allocate space, which is assigned a space, and there is no space for allocation, but there are other operations, examples:
ArrayList Resultlist = new arraylist ();
Try {
Resultlist = someModule.getResultlist ();
} catch (exception e) {
Return NULL;
}
...
Resultlist assigned space at the same time while the declaration, but below with it to point to additional returns. Although we don't have to consider a headache problem such as memory allocation, Java has its own memory management mechanism, but Java is cost-effective, so the waste efficiency of this is. Do you have to avoid it.
3.2 Cycling Excess
We often find out the conditional elements to operate from a array or collection, and often forget to jump out of the cyclic body after execution. example:
String [] Week = new string [] {"Sun", "mon", "tue", "wen", "tu", "fri", "sat"};
For (int i = 0; i IF (Week [i] .Equals ("tue")) { System.out.println ("tue is found"); } } If the above program just finds there is TUE in WEEK, you should jump out of the cyclic body after finding it, the correct way is as follows: For (int i = 0; i IF (Week [i] .Equals ("tue")) { System.out.println ("Tue is Found"); Break; (or possibly Return) } } ... The principle of loop processing is to jump out of the place where it is handled. 3.3 Repeat statements It is recommended that if there are two or more places you need to use the same block code, you should take into account the use of functions, if a function block is independent, it is possible to be called in other calls, this time is also used as usual use. Open. In addition, there is a conditional branch statement, and each branch needs to perform some of the same statement. At this time, you need to mention the outside of the branch, and give a few examples below: Ø How to use Public void someMethod1 () { ... String str = "this is a class"; // For a certain process of STR, return a new STR IF (str! = null) { Str = ... } ... } Public void someMethod2 () { ... String str = "this is a class"; // For a certain process of STR, return a new STR IF (str! = null) { Str = ... } ... } The function of the blue font part can be done using a separate method so that the same method is called in two places. Help in the later maintenance, do not have to go everywhere. Rewriting is as follows: Public void someMethod1 () { ... String str = "this is a class"; Str = GetSomestr (STR); ... } Public void someMethod2 () { ... String str = "this is a class"; Str = GetSomestr (STR); ... } / ** * For a certain process of STR, return a new STR * / Private string getsomester (string str) { IF (str! = null) { Str = ... } Return Str; } Ø Statement merge ... IF (expression1) { ... Somestatement; } else if (expression2) { ... Somestatement; } else { ... Somestatement; } ... or Switch (flag) { Case Result1: ... Somestatement; Break; Case Result2: ... Somestatement; Break; DEFAULT: ... Somestatement; Break; The blue font part is part of each condition branch, which does not have to be written inside each branch, but to adjust the conditional statement uniformly, rewritten as follows: ... IF (expression1) { ... } else if (expression2) { ... } else { ... } Somestatement; ... or Switch (flag) { Case Result1: ... Break; Case Result2: ... Break; DEFAULT: ... Break; } Somestatement; 4. Parameter delivery problem 4.1 General object transfer When calling a method, when the parameter is an object, the default is to pass the address, similar to the pointer inside C / C , Java called "handle". Example: public class testpassparam { Public someclass param = new someclass (); Public void processparam (someclass newparam) { System.out.println (newparam == param); } Public static void main (String [] args) { TestClassparam testclassparam = new testclassparam (); TestClassparam.ProcessParam (TestClassparam.Param); } } The result should return true; because they point to the same handle, that is, the address is the same. If the parameter is the basic data type, it is passed according to the value, and Java does not pass reference to C. Look at the example: Private SomeActionform Processctionform (SomeActionform MyForm) { SomeActionform Otherform = MyForm; Otherform ... .. ...; Return Otherform; } The call is as follows: MyForm = ProcessActionform (MyForm); Among them, MyForm is an object. When it is passed to ProcessActionform, I will pick up with another Otherform. At this time, Otherform also points to myForm's handle, and then use MyForm to pick up the return value, MyForm At this time, it is actually pointing to the original handle. Although the result is correct, there is a problem at this time that there are two alias points to the same handle. Execute the method of any one of these individual names, another alias has changed. So the above example should be rewritten as: Private Void ProcessActionform (SomeActionform myform) { MyForm ... .. ...; Return Otherform; } The call is as follows: Processactionform (MyForm); When the parameters are passed between the screen unless you can guarantee that your data is not tampered with others, you will be careful, or you need to re-clone an object pass, then take the return value. Note: If you pass parameters between EJB's Client and Server, there is no handle pass, and the data is not changed after the server is modified. Because they are converted over the network, there is no condition in which the handle address is the same. You can use this new Object if you need to use this new object if the Client side needs to use this new Object. (EJB2.0 supports Local Interface. In this case, it is possible to pass the same handle, I haven't tried it). 4.2 String's Specialty String is designed in Java to be safe String, and any of String is a String copy and then operates this copy. So String can be seen as a value transfer when the parameter is passed. That is, if you need to modify a string and return to the modified String, you have to pick up the return value. Such as: String str = "this is a name"; Str = Editstr (STR); System.out.println (STR); // "Here is a sample" Private string editstr (String str) { String newstr = Str. Substring (4); Newstr = "here" newstr; Return newstr; } If you want to pass by handle, you can use a class StringBuffer used by String's internal operation, which is made on the same object, so efficiency is naturally high. The above example is rewritten with StringBuffer as follows: StringBuffer str = New StringBuffer ("this is a name"); Editstr (STR); System.out.println (STR); // "Here is a sample" Private void Editstr (StringBuffer Str) { Str.Replace (0, 4, "here"); } 5. Exception treatment Some newcomers are always accustomed to using the return value for error handling. If this method is used, this method will make the program structure more reasonable and more efficient. For example, in the client side, it is necessary to perform DB operation via EJB. The Client side needs to know if the DB processing can pass through the upward Exception method, and the CLIENT end needs to be processed, and then perform exception processing. Such as: Backstage DB processing: Public static java.sql.timestamp getdbsysdate () throws cscWebexception { Connection conn = NULL; Timestamp system = NULL; Try { CONN = pJEJBSVRUTIL.GETWLPOOLCONNECTION (); Systime = comMMondao.getdbsysdate (conn); } catch (SQLEXCEPTION CE) { Throw new cscwebexception (ce.getMessage ()); } finally { Try { IF (conn! = null) { CONN.CLOSE (); } } catch (exception e) { Throw new cscWebexception (E.GetMessage ()); } } Return Systime; } reception: Try { CommONintf.getdbsysdate (); } catch (cscWebexception CEX) { Cat.debug ("", CEX); Return getExceptionForward (CEX); } To prevent violations from being handled, unless you don't need to deal with it, advocate that it will be throw up in the event of Exception, and it is handled by the final call, and certainly cannot be generalized, depending on the situation. For example, I want to solve it in the method in the method, give the bug that appears in a CSC. Public Boolean ChecktelFormat (String Telno) {Boolean Error = False; IF (telno == null || telno.equals ("))) { Error = True; } else { IF (ejb.util.stringutil.chkphone (telno)) { Error = FALSE; } else { Error = True; } } IF (Telno.startSwith) || Telno.startSwith ("186")) { IF (Telno.Length () == 3) { Error = True; } } IF (error) { Objmngr.ShowError ("MCSTC001E"); CMbtelno.requestfocus (); Return False; } else { Return True; } } Used a Flag to remember each Check result, then in the end of Error Dialog, this is a typical C's writing, rewriting as follows: Public Boolean ChecktelFormat (String Telno) { Try { IF (telno == null || telno.equals ("))) { Throw new Exception (); } else { IF (! Ejb.util.stringutil.chkphone (telno)) { Throw new Exception (); } } IF (Telno.startSwith) || Telno.startSwith ("186")) { IF (Telno.Length () == 3) { Throw new Exception (); } } } catch (exception e) { Objmngr.ShowError ("MCSTC001E"); CMbtelno.requestfocus (); Return False; } Return True; } Using the method of throwing Exception, at the end of the method, such an error can be handled immediately, and the efficiency is also the highest. 6. Database operation problem According to existing development experience, usually we take the connection in SessionBean (or Module beans without EJB) and then call the method in which the Database (DAO) is used, that is, the open and closing operation of Connection in the bean, Statement and ResultSet operations in DAO, generally need to be closed at the end. An example of the joint, there is such a method in Commondao: Public static java.sql.timestamp getDBSysdate (Connection Conn) Throws cscwebexception { Statement Stmt = NULL; ResultSet RS = NULL; Timestamp system = NULL; Try { STMT = conn.createstatement (); RS = stmt.executeQuery ("SELECT SYSDATE"); IF (rs.next ()) {systime = rs.getTimeStamp ("sysdate"); } } catch (SQLEXCEPTION CE) { Throw new cscwebexception (ce.getMessage ()); } catch (exception e) { Throw new cscWebexception (E.GetMessage ()); } finally { Try { IF (rs! = null) { Rs.close (); } IF (stmt! = null) { Stmt.close (); } } catch (exception e) { Throw new cscWebexception (E.GetMessage ()); } } Return Systime; } The statement in Finally is always executed, so even the SQLException will be executed even if SQLException is thrown. 7. INDEX Crossing The Index Crossing Includes a lot: Ø String's index String's index starts from 0, the maximum length of its character. Commonly used substring method: String str = "this is a class"; Str.Substring (5, 7); return IS; Str.SUBSTRING (15, 16) or str.substring (15); return last character E; Str.Substring (16, 17); this will throw java.lang.stringindexoutofboundsexception errors; Ø array subscript The array subscript is also started by 0, the maximum length -1, example: Int [] columnlen = new int = {20, 30, 40, 30, 25, 50, 65, 100}; COLUMNLEN [0] is 20; COLUMNLEN [7] is 100; Columnlen [8] will throw java.lang.ArrayindexOutofboundsexception errors. Ø Vector, ArrayList, etc. Collection Size Vector and ArrayList belong to list, they are all ordered, the subscript starts from 0, and the maximum length is -1, and the array is different, their elements must be Object, but can be different types Object, but after the removal, the type conversion is to be performed. The array has the same type of all elements. When the ArrayList is constructed, if there is no element, if the set (INDEX, OBJECT) will be wrong. Before you have to do Add (Object). example: ArrayList List = New ArrayList (10); // This place is only the initial capacity of the List, does not represent 10 elements, this is different from the array, the array has 10 initial values in this case, initial The value is related to the specific element type, the general object is null; List.set (0, "first"); // will throw java.lang.indexOutofboundsexception List.add ("first"); List.set (0, "new first"); // is correct, because the location 0 already exists 8. other Specific to each item, different API usage may encounter some common problems. It is necessary to carry out the necessary training at the beginning of the project, often give a good Sample Half of the SAMPLE. Also use some problems here no longer detailed description: ü Size write problem (String.equals () and string.equalsignorecase (), variable name, etc.) Ü Braces matching problem ü. . . 9. Problem to be discussed 9.1 Method Returns place private int getStatus (String sFlag) {if (CodeBook.STATUS_OK_NAME.equals (sFlag)) {return CodeBook.STATUS_OK_VALUE;} else if (CodeBook.STATUS_OK1_NAME.equals (sFlag)) {return CodeBook.STATUS_OK1_VALUE;} else if (CodeBook. Status_ng_name.equals (sflag)) {return codebook.status_ng_value;} Return codebook.status_ng_value;} versus Private int getStatus (string sflag) { INT ISTATUS = Codebook.status_ng_value; IF (codebook.status_ok_name.equals (sflag)) { ISTATUS = Codebook.status_ok_value; } else if (codebook.status_ok1_name.equals (sflag)) { ISTATUS = Codebook.status_ok1_value; } else if (codebook.status_ng_name.equals (sflag)) { iStatus = codebook.status_ng_value; } Return ISTATUS; } 9.2 Variable Declaration Point example: For (int i = 0; i <10; i ) { String stmp = string.valueof (i); ... } versus String stmp = null; For (int i = 0; i <10; i ) { STMP = String.Valueof (i); ... }
