Files

The game that remembered

• Until 1987, switching off a games console erased everything. Every session started from the first screen; a long game was a long afternoon or nothing.
• The Legend of Zelda shipped with a small battery inside the cartridge that kept a few hundred bytes alive. For the first time a home game could save, and pick up where you left off.
• Every high score, every document, every setting you have ever kept exists for the same reason: RAM forgets when the power goes, and a file 文件 does not.
• This lesson is why files exist and how pseudocode reads and writes a text file, line by line.

Why files

• Variables live in RAM and disappear when the program ends. A file on secondary storage 辅助存储器 keeps data between program runs: persistence 持久性.
• Files let one program's output become another program's input, so data can be shared.
• A file can hold more than fits in memory, and a program can restart from a saved state.

The variable is gone at the end of the run; the file is still there tomorrow

Text files and the three modes

• A text file 文本文件 stores data as lines of characters, which a program reads and writes one line at a time, from the start. That sequence of lines is what lets a loop work through it.
• OPENFILE … FOR READ reads from the first line. FOR WRITE creates a new file, or overwrites an existing one. FOR APPEND adds new lines at the end, keeping what is there.
• Open before use, close after. A file has one mode at a time.

Reading every line

OPENFILE "data.txt" FOR READ
WHILE NOT EOF("data.txt") DO
    READFILE "data.txt", LineString
    OUTPUT LineString
ENDWHILE
CLOSEFILE "data.txt"

• EOF tests for the end of file 文件结束. Test it before each read: reading past the end is an error.
• CLOSEFILE matters: buffered data may be lost and other programs locked out of a file left open.

Writing lines

OPENFILE "log.txt" FOR WRITE
FOR i ← 1 TO 100
    WRITEFILE "log.txt", "Event " & NUM_TO_STR(i)
NEXT i
CLOSEFILE "log.txt"

• WRITE starts the file empty. To add to a log that must keep its history, open it FOR APPEND instead.
• A text file holds strings: a number is converted with NUM_TO_STR on the way out.

Worked example: the last three lines of a file

PROCEDURE LastLines(BYVAL FileName : STRING)
    DECLARE LineX, LineY, LineZ : STRING
    LineX ← ""
    LineY ← ""
    LineZ ← ""
    OPENFILE FileName FOR READ
    WHILE NOT EOF(FileName) DO
        LineX ← LineY
        LineY ← LineZ
        READFILE FileName, LineZ
    ENDWHILE
    CLOSEFILE FileName
    OUTPUT LineX
    OUTPUT LineY
    OUTPUT LineZ
ENDPROCEDURE

Fields in a line

• A record is written as one line, its fields joined by a separator 分隔符 character, and each number or Boolean converted with NUM_TO_STR (and read back with STR_TO_NUM, or by comparing with "TRUE").
• Choose a separator that can never appear in the data: a comma or | for names and numbers; never a space when a name may contain one.
• Reading back: read the line, split it at the separator, convert each field to its type, store it in the record.

Join with a separator to write; split and convert to read

Worked example: save a record, load it back

OPENFILE "scores.txt" FOR APPEND
WRITEFILE "scores.txt", Player.Name & "," & NUM_TO_STR(Player.Score)
CLOSEFILE "scores.txt"

OPENFILE "scores.txt" FOR READ
READFILE "scores.txt", Line
Pos ← POSITION(Line, ",")
Player.Name ← LEFT(Line, Pos - 1)
Player.Score ← STR_TO_NUM(MID(Line, Pos + 1, LENGTH(Line) - Pos))
CLOSEFILE "scores.txt"

Serial and direct access

• A text file is serial 顺序 (sequential): it is read in order from the start, so finding one line means reading every line before it.
• Direct (random) access jumps straight to a record at a calculated position without reading the others; it needs fixed-length records, which text files do not have.
• For the syllabus, the text-file loop above is the model: open, test EOF, read, close.

Marks that slip away

You've got it