postgresql - 使用批处理文件将 csv 的第一行解析为 sql 表

Question

我有一些csv文件，我需要打开csv文件，读取csv的第一行并将其转换为临时sql表，然后将数据加载到sql表中，如下所示:

读取 CSV 的行，对于每一行:
把它分解成字段创建一个临时sql表将这些字段插入到数据库表的一行中

我试过这样的东西

这个脚本现在分为4个部分，文件初始化；文件创建、处理和复制数据， 一切正常，除了在 fil.sql 上我得到的输出为

                       CREATE TEMP TABLE temtab(
                        firstcolumn character varying (255),
                         secondcolumn character varying (255),
                            lastcolumn character varying (255),
                            );
                         \COPY temtab from bio.csv WITH DELIMITER ; csv HEADER

虽然我希望最后一列没有逗号

                       CREATE TEMP TABLE temtab (
                       firstcolumn character varying (255),
                       secondcolumn character varying (255),
                         lastcolumn character varying (255)
                           );
                   \COPY temtab from bio.csv WITH DELIMITER ; csv HEADER







            @echo off
            ::setlocal enabledelayedexpansion
            REM Assiging dir to current directory
              SET dir=%CD%
              REM Defining database name
               SET dbname=****
               REM Defining Host name
                 SET host=****
               REM Defining user
                 SET user=****
                 REM Defining Port
         SET port=****
               REM SQL file where query is to be executed
                 SET sqfile=fil.sql

           SET fi=bio.csv
                 call:fileinitialization

                call:filecreation 

                call:proces

                  call:copydata
            goto:eof

         :fileinitialization
          REM Assigning name of temporary table 
                  SET tabnam=temtab
             REM  Setting delimiter to variable delim
             SET delim=;
        REM Declaring variable numfields to store index of variable names array  
    set numFields=0
    echo para setted
    set fi=bio.csv
    SET tex=text
    SET com=,
                GOTO:EOF

           :filecreation 
          REM Setting create temporary table command with table name tabnam
             SET creat=CREATE TEMP TABLE %tabnam%
             echo %creat%

                     GOTO:EOF 

                :proces
                REM Executing loop for each file in current directory
          echo %creat%>fil.sql
        REM Read the lines of the CSV file
        For /F  "eol==" %%A in (bio.csv) Do ( set "line=%%A" 


                REM check if index of array is 0
                     if !numFields! equ 0  (
                      REM Fisrt line, Store in array name
                         for %%B in (!line: ^=!) do (
                echo %%B character varying (255^),>>fil.sql   

                        set /A numFields+=1
                    set name[!numFields!]=%%B
                    ) ) )





            GOTO:EOF

             :copydata
           echo \COPY %tabnam% from %fi% WITH DELIMITER %delim% csv HEADER
            echo \COPY %tabnam% from %fi% WITH DELIMITER %delim% csv HEADER;>>fil.sql
                  GOTO:EOF  
              ::endlocal
                     Pause

Answer 1

虽然我不知道 SQL 表的格式，但我可以向您展示如何读取 CSV 文件。下面的批处理文件从文件中读取所有行；它首先从第一行(CSV 标题)获取字段名称并创建一个变量名称的数组(消除字段名称中可能的空格)；然后它读取其余行并将每个字段值分配给其相应的 Batch 变量。

处理CSV.BAT:

@echo off
rem General-purpose CSV file reader program
rem Antonio Perez Ayala

setlocal EnableDelayedExpansion
set numFields=0
rem Read the lines of the CSV file
for /F "delims=" %%a in (CSVfile.csv) do (
   set "line=%%a"
   if !numFields! equ 0 (
      rem It is the first line: break it into an array of field names (removing spaces)
      for %%b in (!line: ^=!) do (
         set /A numFields+=1
         set name[!numFields!]=%%b
      )
   ) else (
      rem Replace spaces by Ascii-128 (to avoid split values that may have spaces)
      set "line=!line: =Ç!"
      rem Insert any char. at beginning of each field, and separate fields with spaces
      set i=0
      for %%b in (X!line:^,^= X!) do (
         set "field=%%b"
         rem Recover spaces in this field, if any
         set "field=!field:Ç= !"
         rem And assign it to corresponding variable (removing first character)
         set /A i+=1
         for %%i in (!i!) do set "!name[%%i]!=!field:~1!"
      )
      rem At this point all variables have the values of current record.
      rem They may be accessed explicitly (ie, from example CSVfile.csv):
      echo/
      echo Record of !FirstName! !LastName!
      rem ... or implicilty via the NAME array:
      for /L %%i in (3,1,!numFields!) do (
         for %%b in (!name[%%i]!) do echo    %%b: !%%b!
      )
   )
)

CSV文件.csv:

First Name,Last Name,Address,Postal Code,Company,Departament,Floor,Phone,Mobile
John,Smith,123 Fake Street,45612,SomeCo,Accounting,4,123-555-5555,123-555-5556
Jane,Doe,123 Fake Street,,SomeCo,,4,123-555-5555,123-555-5556

输出:

Record of John Smith
   Address: 123 Fake Street
   PostalCode: 45612
   Company: SomeCo
   Departament: Accounting
   Floor: 4
   Phone: 123-555-5555
   Mobile: 123-555-5556

Record of Jane Doe
   Address: 123 Fake Street
   PostalCode:
   Company: SomeCo
   Departament:
   Floor: 4
   Phone: 123-555-5555
   Mobile: 123-555-5556

请注意，此程序使用了多种高级批处理技术。我建议您对每个您不完全理解的命令(即:SET/?)寻求帮助并仔细阅读。如果在这个过程之后您对这个程序还有其他问题，只需将它们作为编辑发布到您的原始问题中。

这个程序最复杂的部分是负责在相应字段为空时(两个逗号并排)将空字符串赋值给变量；如果文件没有空字段，程序可能会更简单一些。此外，如果文件中出现某些特殊的 Batch 字符，如 !，此程序(与大多数 Batch 解决方案一样)可能会给出错误的结果。如果需要，可以通过程序中的某些修改来管理这些字符中的大多数。

编辑: 不存在空字段时的修改版本

@echo off
rem CSV file reader program when no empty fields exist
rem Antonio Perez Ayala

setlocal EnableDelayedExpansion
set numFields=0
rem Read the lines of the CSV file
for /F "delims=" %%a in (CSVfile.csv) do (
   set "line=%%a"
   if !numFields! equ 0 (
      rem It is the first line: break it into an array of field names (removing spaces)
      for %%b in (!line: ^=!) do (
         set /A numFields+=1
         set name[!numFields!]=%%b
      )
   ) else (
      rem Replace spaces by Ascii-128 (to avoid split values that may have spaces)
      set "line=!line: =Ç!"
      rem Separate fields (using comma as standard Batch separator)
      set i=0
      for %%b in (!line!) do (
         set "field=%%b"
         rem Assign this field to corresponding variable, recovering spaces
         set /A i+=1
         for %%i in (!i!) do set "!name[%%i]!=!field:Ç= !"
      )
      rem At this point all variables have the values of current record.
      rem They may be accessed explicitly (ie, from example CSVfile.csv):
      echo/
      echo Record of !FirstName! !LastName!
      rem ... or implicilty via the NAME array:
      for /L %%i in (3,1,!numFields!) do (
         for %%b in (!name[%%i]!) do echo    %%b: !%%b!
      )
   )
)

请注意，除了空格之外，FOR 集合中的标准分隔符是逗号、分号和等号:

for %a in (one two,three;four=five) do echo %a

以前的程序用另一个字符替换空格并使用逗号分隔字段。但是，如果该行可能包含分号或等号，则字段将在该点被拆分，因此在这种情况下，必须在 FOR 之前将这些字符更改为另一个字符，然后以与空格相同的方式恢复。

编辑: 新请求的修改(删除最后一个逗号)

消除最后一个逗号并不简单，但也不是太复杂。我希望我的方法容易理解；它基于显示文本(输入提示)的 SET/P 命令行为，末尾没有新行；请注意，格式为 SET /P =text>>out<NUL 。 <NUL 部分是必需的，因此 SET/P 不会等待输入；不要在 < 前留空格(与 >> 相同)。但是，我认为这种行为在 Windows Vista 更高版本中不起作用。如果该方法对您不起作用，则必须再次修改...

我也继续前进，包括一些关于您的代码中仍然缺少的部分(我认为)的评论，即几个文件的处理。

:proces
REM Executing loop for each file in current directory
REM *This may be done with a FOR loop:*
::*for %%F in (*.csv) do (*
    REM *The file name is given by %%F. In this case, the fileinitialization part*
    REM *must be done here, for example:*
    set numFields=0
    echo %creat%>fil.sql
    REM Read the lines of the CSV file
    For /F "eol==" %%A in (bio.csv) Do ( 
        set "line=%%A" 
        REM check if index of array is 0
        if !numFields! equ 0  (
            REM First line, Store in array name
            for %%B in (!line: ^=!) do (
                REM Note that I changed the place of the ECHO command
                set /A numFields+=1
                set name[!numFields!]=%%B
                if !numFields! equ 1 (
                    REM First field: show it with NO comma and NO NEW LINE
                    set /P =%%B (text^)>>%sqfile%<NUL
                ) else (
                    REM Next fields: complete the comma of previous field, WITH NEW LINE
                    echo ,>>%sqfile%
                    REM ... and show this field with NO comma and NO NEW LINE (again)
                    set /P =%%B (text^)>>%sqfile%<NUL
                )
            )
            REM Insert the new line of last field (that have NOT comma :-)
            echo/>>%sqfile%
        )
    )
::*)*
GOTO:EOF

:copydata

我强烈建议您保留我以前的格式:每个代码块内有 4 个对齐列，用括号括起来，并将右括号放在开始命令 FOR 或 IF 的同一列中。这种格式将帮助您轻松定位大型程序中因括号不匹配而导致的错误。

安东尼奥