concatenate text files into 1 csv and import it to SQL database

[Cyrus87]

Hello,

I would like to concatenate thousands .txt files to 1 .CSV file and import it to a SQL database.

All txt files have the same structure.

I just want bold content of the txt file:

boucham;20150521;ECOM;00066608CECF;ZEBRA;AC3FA415C315;IMEI;355282030665496;IMSI;208014802943276;WinCE;5.2.23137;232.73;159.07;SIEMENS/HC25;REVISION

My Table:

 

Could you help me to create the script, this is my first in autoit.

I have this at the moment

$constrim="DRIVER={SQL Server};SERVER=;DATABASE=importtxt;uid=;pwd=;"
$adCN = ObjCreate ("ADODB.Connection") ; <== Create SQL connection
$adCN.Open ($constrim) ; <== Connect with required credentials
MsgBox(0,"",$constrim )

if @error Then
    MsgBox(0, "ERROR", "Failed to connect to the database")
    Exit
Else
    MsgBox(0, "Success!", "Connection to database successful!")
EndIf

 

Edited May 24, 2015 by Cyrus87

[water]

Why create one large CSV file first?
Connect to the database and then import all the CSV files in a loop.

Edited May 24, 2015 by water

[Cyrus87]

i thought it could accelerate the import

[water]

Should the import script crash for whatever reason then restarting with the first unprocessed file is much easier with the approach I suggest.
Do the import of each CSV file in a transaction and commit when the import was successful. Write the name of the CSV file to an INI file when you start the import and delete whenever the import was successful. So you know where to restart.

[mikell]

water,
jchd advised him to concatenate the csv files to accelerate the import on the french forum

[water]

Thanks for the info.
I do not watch this forum because my French is very bad (at least I was told so by my teacher).

I think the "best" solution depends on many factors:

• Is this a one time thing or does the import need to be run regularly?
• Is speed important?
• Should it be possible to restart the import after x files?
• etc.

Would be great to get some info from the OP.

 

 

[AspirinJunkie]

I think it will be faster if you use the ".import"-command by using _SQLite_SQLiteExe().

[water]

On the french forum he is talking about "SQL server 2005 (9.0.3042)" he wants to import the data to.
Is SQLite an option then?

[AspirinJunkie]

You're right. I did'nt read carefully.

Maybe a general hint: collect all insert queries and execute them at once. Maybe the use of begin/commit can speed up a little bit when it's leading to a avoid of many single HDD writes. 

[jchd]

The main reason why I advised to group thousands files first was that, in case of error it will be much easier to spot it, fix it, drop the target table and re-launch the import as a whole. The OP also said he wasn't expecting errors in input CSV data.

I have zero experience with SQL server but I bet it offers tools to import CSV data easily.

Last point, the dates in input are in format YYYYMMDD and I suspect it would need to first regexp the whole input CSV to format them YYYY-MM-DD (ISO) as it is more likely to cope with allowable date format for SQL server CSV import.

[Cyrus87]

I would like to run it one time in the week to update the database.

Speed is not important

Should i use FileReadToArray function to import txt files ?