/*
README for FoxDbf java source *
FOXPRO databases in java. Supports all basic field types: C,N,F,D,L,I,B,M,G,T,Y. Partly supports IDX,CDX index, bugs inside. Writing .idx on Your risk, really corrupts index files. Set READ_ONLY to be safe not to lose data/indexes when only data reading needed.
db = new base(); db.Field[x] contain structures of fields and data of the current record (in memory, debug values)
db.eof, db.bof, db.reccount, db.recno, db.deleted - the same as functions in FoxPro
Functions similar to FoxPro commands:
use(
additions:
readBinaryFromFile(
Sample1: to create a database
db.create("SAMPLE1.DBF", "Npk N(10), Name C(20), Rating F(5,2), Birth D, Notes M, ieee B(8,4), Id I, Salary Y"); db.Field[0].setByLong(1); db.Field[1].setByString("Mickey"); db.Field[2].setByDouble( 3.15 ); db.Field[3].setByString("20050328"); db.Field[4].setByString("Some notes about."); db.Field[5].setByDouble( -9.1678 ); db.Field[6].setByLong( -2999999 ); db.Field[7].setByDouble( 3000.25 ); db.insert_from_memory(); // replace all fields in database
int r = db.FieldI("RATING"); //=2 db.Field[r].setByDouble( 6.25 ); db.replace_field_from_memory(r); // replace particular field db.use();
Sample2: to scan database
db.READ_ONLY = true; // do not modify db.use("SAMPLE1.DBF"); while(!db.eof) { long npk = db.Field[ db.FieldI("Npk") ].longValue(); String Name = db.Field[ db.FieldI("Name") ].stringValue(); double Rating = db.Field[ db.FieldI("Rating") ].doubleValue(); String Notes = db.Field[ db.FieldI("Notes") ].stringValue(); String Birth = db.Field[ db.FieldI("Birth") ].stringValue(); //"YYYYMMDD" double Salary = db.Field[ db.FieldI("Salary") ].doubleValue(); db.skip(1); } db.use(); db.READ_ONLY = false; // writing enabled
Sample3: DOS table
db.DOS_FOX = true; db.create("SAMPLE3.DBF","MyMemo M, Logic L, Birth D"); db.use();
Sample4: Binary data, or bytes to be precise for special or regional characters in C,M fields (data are byte[] elements)
db.BINARY_CHARS = true; // for G-type General fields binary mode is by default db.create("SAMPLE4.DBF", "Name C(10), Picture M"); // store spec.chars in name field byte[] bArray = { 8, 'M','A',13, 0, 1 }; db.Field[ db.FieldI("Name") ].setBinaries( bArray ); // save picture file to memo field (or G-type) byte[] bPicArray = db.readBinaryFromFile( "profile.jpg" ); db.Field[ db.FieldI("Picture") ].setBinaries( bPicArray ); db.append_blank(); db.replace_from_memory();
// create a new picture file from memo field byte[] bPicArray2 = db.Field[ db.FieldI("Picture") ].Binaries(); db.writeBinaryToFile( "profile2.jpg", bPicArray2 ); db.use();
WORKING WITH INDEXES
This code is good enough when working without indexes, or if You use indexes without replacing/inserting new data. Really really, do not take it as a database engine. This is java, indexing algorithms are too complex and poorly described, so this simplified java may corrupt the index files.
Anyway, this code is good to search in ordered database by key and scan records. If writing in indexes required then consider using original FoxPro process in background or try using harbour project http://harbour-project.org/
Functions similar to FoxPro commands:
IDX file:
set_order_to_idx(< .idx file name>) - sets index file, similar SET INDEX TO
create_idx(< .idx file name>,
CDX file (read only):
set_cdx_order(
SEEK() - does searching according to data in memory variables Field[x]
The PREPARE_KEY_FROM_DATA() procedure prepares searchKey according to data and searching expression (if key is a field value, it converts automatically, otherwise write java code there to prepare user defined keys as DTOS(data)+.. or SYS(15,data..) , this is not original FoxPro :)
Index sample1: create and write indexed database (prepare for lags)
db.create("/I.DBF", "Numb N(10)"); db.create_idx("/I.IDX","Numb"); // INDEX ON Numb without re-indexing records for(int k=0;k<100;k++) // append 100 indexed records { db.Field[0].setByLong(k); db.insert_from_memory(); } for(int k=99;k>=0;k--) // find them all by key and replace with (5000-value) { db.Field[0].setByLong(k); if(db.SEEK()) { db.Field[0].setByLong(5000-k); db.replace_from_memory(); } else { System.out.print("Not found?! Should be.\n"); // db.Idx.sRecno contains nearest record, so go to this and skip } }
for(int k=4901;k<=5000;k++) // find new values to verify index
{
db.Field[0].setByLong(k);
db.SEEK();
if(!db.eof) System.out.print("Seek " + String.valueOf(k) + ", record " + String.valueOf(db.recno) + "\n");
else System.out.print("Not found?! Should be.\n");
}
db.use();
Index sample2: use database for scanning by using .idx index
db.READ_ONLY = true; // to be sure for reading only db.use("/I.DBF"); db.set_order_to_idx("/I.IDX"); db.go_top(); // scan them sorted all and print values and record numbers while(!db.eof) // should be numbers in descending order { long V = db.Field[ db.FieldI("Numb") ].longValue(); System.out.print( "Numb = " + String.valueOf(V) + ", record " + String.valueOf(db.recno) + "\n"); db.skip(1); } db.use(); db.READ_ONLY = false; // set writing mode on
// Verify in FoxPro by commands // USE C:\I // SET INDEX TO C:\I.IDX // BROW
Index sample3: use database for scanning by using .cdx index Prints all "Bob" records
db.use("/MyBASE.DBF"); // this opens also MyBASE.CDX and sets read-only mode db.set_cdx_order("name");
db.Field[db.FieldI("Name")].setByString("Bob"); db.SEEK(); while(!db.eof) // should be numbers in descending order { String fName = db.Field[ db.FieldI("FullName") ].stringValue(); if(!fName.startsWith("Bob")) break; System.out.print( "Full name is " + fName + "\n"); db.skip(1); } db.use();
*/
/* REMARKS 03.2022, noobies faq.
A few notes about real FoxPro and history. FoxPro was a good fast and very cheap "as a database" solution for a PC, or small network. PCs were insufficient, awful quality. While properly designed expensive databases required SQL - with power reliable server responding to multi-users queries. So, FoxPro was used as a temporary prototype database that grew up to a large system due to "no time to rewrite it all". As DBFs,FPT,CDX,IDX files were buffered (wr./read) on each very slow! PC those days by kinda locking-ethernet-hack, there always were power-supply cuts, bad disk drives, lost networks, that corrupted almost everything. Errors like "Not a database file", "Index file does not match database", and similar, like time-bombs blasted and forced the database admin to run and pack DBFs, and reindex, use smart hack tools to restore the database system. Especially with dozens of users waiting. Never again.
Nevertheless the DBF indexing (CDX complex) worked almost instantly, very fast. DBF could reach 2GB size limit. And FPT another 2GB. That was huge amount of data then. It is like a fast JSON request in a read only large table. Very powerful thing. Binary search and not a full-scan. Opens the file and reads data the right way. But no writing, this requires a proper database engine. Or write a new one transaction safe.
The other well working solution, especially serving web pages, was a foxpro process running in background on server that processed appearing request files in folder, and wrote answers to files for php-servers. This is still reliable. Anyway there are NoSQL, EasySQL, MySQL, others nowadays. No big reason to revive the dead. And btw. the UniCode characters, avatar pictures, video streaming, FoxPro is inappropriate for this. Prior to FoxPro, there were DBase and Clipper. The DBF part is similar, so can use the same data tools.
The CDX file writing part is possible to develop, but then this won't work on FoxPro anymore for sure. Now a large indexed database with multiple indexes can be made on FoxPro and used for seeking - fast data lookup.
*/