You can use ADO (ODBC/JET/OLEDB Text Driver) to treat 'decent' .txt/.csv/.tab/.flr files as tables in a SQL Database from every COM-enabled language. Then the comparisons could be done using the power of SQL (DISTINCT, GROUP, (LEFT) JOINS, ...).
Added with regard to your comment:
It's your problem and I don't want to push you where you don't want to go. But SQL is a good (the best?) tool, if you need to compare tabular data. As evidence the output of a script that spots the differences in two .txt files:
======= The .txt files to play with
------- file1.txt
"AC";"AM"
40000;-19083,00
40100;20000,00
40200;350004,00
40300;3498,99
------- file2.txt
"AC";"AM"
40000;-19083,00
40300;3498,99
40105;-234567,00
40200;350,00
======= Some diagnostic SQL
------- <NULL> indicates: In F1 but not in F2 (LEFT JOIN)
SELECT T1.AC, T1.AM, T2.AM FROM [file1.txt] AS T1 LEFT JOIN [file2.txt] AS T2 ON (T1.AC =
T2.AC)
------- Result
AC File1 File2
40000 -19083 -19083
40100 20000 <NULL>
40200 350004 350
40300 3498,99 3498,99
------- <NULL> indicates: Not in the other file (LEFT JOIN, UNION)
SELECT T1.AC, T1.AM, T2.AM FROM [file1.txt] AS T1 LEFT JOIN [file2.txt] AS T2 ON (T1.AC =
T2.AC) UNION SELECT T2.AC, T1.AM, T2.AM FROM [file2.txt] AS T2 LEFT JOIN [file1.txt] AS T1
ON (T1.AC = T2.AC)
------- Result
AC File1 File2
40000 -19083 -19083
40100 20000 <NULL>
40105 <NULL> -234567
40200 350004 350
40300 3498,99 3498,99
------- the problems: missing, different values
SELECT T1.AC, T1.AM, T2.AM FROM [file1.txt] AS T1 LEFT JOIN [file2.txt] AS T2 ON (T1.AC =
T2.AC) WHERE T2.AM IS NULL OR T1.AM <> T2.AM UNION SELECT T2.AC, T1.AM, T2.AM FROM [file2.
txt] AS T2 LEFT JOIN [file1.txt] AS T1 ON (T1.AC = T2.AC) WHERE T1.AM IS NULL OR T1.AM <>
T2.AM
------- Result
AC File1 File2
40100 20000 <NULL>
40105 <NULL> -234567
40200 350004 350
Further additions:
This article deals with ADO and text files; look for a file adoNNN.chm
(NNN=Version number, e.g. 210) on your computer; this is a good book about
ADO.
You can use Access or OpenOffice Base to experiment with SQL statements
applied to a linked/referenced (not imported!) text database.
A script/program will be easy after you mastered the initial hurdle: connecting
to the the database, i.e. to a folder containing the files and a schema.ini
file to define the structure of the files=tables.
The output above was generated by:
Const adClipString = 2
Dim oFS : Set oFS = CreateObject( "Scripting.FileSystemObject" )
Dim sDir : sDir = oFS.GetAbsolutePathName( ".\txt" )
Dim oDB : Set oDb = CreateObject( "ADODB.Connection" )
oDB.Open "Provider=Microsoft.Jet.OLEDB.4.0;Data Source=" & sDir & ";Extended Properties=""text"""
Dim sSQL
Dim sFiNa
WScript.Echo "=======", "The .txt files to play with"
For Each sFiNa In Array( "file1.txt", "file2.txt" )
WScript.Echo "-------", sFiNa
WScript.Echo oFS.OpenTextFile( "txt\" & sFiNa ).ReadAll()
Next
WScript.Echo "=======", "Some diagnostic SQL"
Dim aSQL
For Each aSQL In Array( _
Array( "<NULL> indicates: In F1 but not in F2 (LEFT JOIN)" _
, Join( Array( _
"SELECT T1.AC, T1.AM, T2.AM FROM" _
, "[file1.txt] AS T1" _
, "LEFT JOIN [file2.txt] AS T2 ON (T1.AC = T2.AC)" _
), " " ) ) _
, Array( "<NULL> indicates: Not in the other file (LEFT JOIN, UNION)" _
, Join( Array( _
"SELECT T1.AC, T1.AM, T2.AM FROM" _
, "[file1.txt] AS T1" _
, "LEFT JOIN [file2.txt] AS T2 ON (T1.AC = T2.AC)" _
, "UNION" _
, "SELECT T2.AC, T1.AM, T2.AM FROM" _
, "[file2.txt] AS T2" _
, "LEFT JOIN [file1.txt] AS T1 ON (T1.AC = T2.AC)" _
), " " ) ) _
, Array( "the problems: missing, different value" _
, Join( Array( _
"SELECT T1.AC, T1.AM, T2.AM FROM" _
, "[file1.txt] AS T1" _
, "LEFT JOIN [file2.txt] AS T2 ON (T1.AC = T2.AC)" _
, "WHERE T2.AM IS NULL OR T1.AM <> T2.AM" _
, "UNION" _
, "SELECT T2.AC, T1.AM, T2.AM FROM" _
, "[file2.txt] AS T2" _
, "LEFT JOIN [file1.txt] AS T1 ON (T1.AC = T2.AC)" _
, "WHERE T1.AM IS NULL OR T1.AM <> T2.AM" _
), " " ) ) _
)
sSQL = aSQL( 1 )
WScript.Echo "-------", aSQL( 0 )
WScript.Echo sSQL
Dim oRS : Set oRS = oDB.Execute( sSQL )
WScript.Echo "------- Result"
WScript.Echo Join( Array( "AC", "File1", "File2" ), vbTab )
WScript.Echo oRS.GetString( adClipString, , vbTab, vbCrLf, "<NULL>" )
Next
oDB.Close
If you delete/ignore the fat (create SQL statements, diagnostics output), it boils
down to 6 lines
Dim oDB : Set oDb = CreateObject( "ADODB.Connection" )
oDB.Open "Provider=Microsoft.Jet.OLEDB.4.0;Data Source=" & sDir & ";Extended Properties=""text"""
sSQL = "..."
Dim oRS : Set oRS = oDB.Execute( sSQL )
WScript.Echo oRS.GetString( adClipString, , vbTab, vbCrLf, "<NULL>" )
oDB.Close
which can be 'ported' easily to every COM-enabled language, because the ADO
objects do all the heavy lifting. The .GetString method comes handy, when you
want to save a resultset: just twiddle the separator/delimiter/Null arguments
and dump it to file
oFS.CreateTextFile( ... ).WriteLine oRS.GetString( _
adClipString, , ",", vbCrLf, ""
)
(don't forget to add a definition for that table to your schema.ini). Of
course you also can use a "SELECT/INSERT INTO", but such statements may not
be easy to get right/passed the ADO Text Driver's parser.
Addition wrt Computations:
Start with a 5 x 2 master/approved file containing:
Num0 Num1 Num2 Num3 Num4
7,6 6,1 3,8 0,9 8,9
0,9 9,4 4,7 8,8 9,9
transform it to expected.txt
Num0 Num1 Num2 Num3 Num4 Spot
7,6 6,1 3,8 0,9 8,9 1
0,9 9,4 4,7 8,8 9,9 2
by appending the Spot column so it conforms to
[expected.txt]
ColNameHeader=True
CharacterSet=1252
Format=Delimited(;)
Col1=Num0 Float
Col2=Num1 Float
Col3=Num2 Float
Col4=Num3 Float
Col5=Num4 Float
Col6=Spot Integer
in your schema.ini file. Similarly, transform a measure file like:
Num0 Num1 Num2 Num3 Num4
7,1 1,1 3,8 0,9 8,9
0,9 9,4 4,7 8,8 9,9
to measured.txt
Num0 Num1 Num2 Num3 Num4 Spot
7,1 1,1 3,8 0,9 8,9 1
0,9 9,4 4,7 8,8 9,9 2
Apply
sSQL = Join( Array( _
"SELECT E.Num0 - M.Num0 AS Dif0" _
, ", E.Num1 - M.Num1 AS Dif1" _
, ", E.Num2 - M.Num2 AS Dif2" _
, ", E.Num3 - M.Num3 AS Dif3" _
, ", E.Num4 - M.Num4 AS Dif4" _
, ", E.Spot AS Spot" _
, "FROM [expected.txt] AS E" _
, "INNER JOIN [measured.txt] AS M" _
, "ON E.Spot = M.Spot" _
), " " )
Write the resultset to differences.txt
aFNames = Array( "Num0", ... "Spot" )
oFS.CreateTextFile( sFSpec ).Write _
Join( aFNames, sFSep ) & sRSep & oRS.GetString( adClipString, , sFSep, sRSep, "" )
and you get:
Num0 Num1 Num2 Num3 Num4 Spot
0,5 5 0 0 0 1
0 0 0 0 0 2
