All Forums Analytics
Sankalp.C 45 posts Joined 08/13
16 Dec 2013
Fast Load Vs Multiload

In case of ideal senario for FastLoad why do we consider Fastload instead of Multiload where both are Bulkload(64kb).
Correct me if I am wrong, the populare answer I got is is Fast Load use 2 phaze where Multiload consists of 5, which I really doubt.
Cause the  phase of Multiload -
- Initialization of Layout
- DML
- Acquisition
- Application
- Clean off
all are this are applicable for even fast load aren't they?
 
 

Raja_KT 1246 posts Joined 07/09
16 Dec 2013

Fastload has two phases: acquisition phase and application phase. mload has 5 phases(Note: however there is no acquisition phase for mload delete).
mload phases: Preliminary,DML Transaction,Acquisition,Application,Cleanup.
The name itself is fast load.
Cheers,
Raja
 

Raja K Thaw
My wiki: http://en.wikipedia.org/wiki/User:Kt_raj1
Street Children suffer not by their fault. We can help them if we want.

M.Saeed Khurram 544 posts Joined 09/12
16 Dec 2013

Hi San,
In your first line "In case of ideal senario for FastLoad why do we consider Fastload instead of Multiload" You know that the scenario is ideal for fastload, then what made you think about MLOAD for this scenario? 
 

Khurram

ulrich 816 posts Joined 09/09
16 Dec 2013

Not sure that above answers the orig question.
I guess the number of phases are irrelevant. Usually total time to get things done is what matters.
But both utilites come with limitations.
Fastload can only load into empty tables.
Multiload can also load data into populated tables but support only NUSIs on the table.
Multiload loads the data into work tables. Prepares the data block wise and apply whole block changes.
These days Fastload with NOPI and merge is comparable fast but will have less limitations on indices and can also has more benefits. Some customer are using this as default szenario and I guess number is growthing.
Considering "ideal senario" which means empyt table. The resource and time difference is not very big. Automated NUSI maintenance is easier with MLOAD as it does not require separate steps. Also you might want to deal with less utilities as they all have different syntax and debuging needs. So if you use MLOAD for some cases it might be worth to run all loads with MLOAD only.
 

feel free to donate bitcoin:12kgAUHFUqvG2sQgaRBXFhCwyf9HXdkGud

M.Saeed Khurram 544 posts Joined 09/12
16 Dec 2013

Hi San, 
Sorry if my question is not clear, In fact you wrote that the scenario is ideal for Fastload, thats why I deduced you know much about these utilities. 
Ulrich has explained what I was thinking. Hope you got answered.
 

Khurram

Sankalp.C 45 posts Joined 08/13
16 Dec 2013

Thanks! Rajat.
Thanks! Ulrich.
Thanks! khurram.

Raja_KT 1246 posts Joined 07/09
16 Dec 2013

Hi Sankalp,
You can address me as Raja (not Rajat...it has different meaning :)).
Cheers,
Raja

Raja K Thaw
My wiki: http://en.wikipedia.org/wiki/User:Kt_raj1
Street Children suffer not by their fault. We can help them if we want.

Raja_KT 1246 posts Joined 07/09
16 Dec 2013

I feel that at the end of the day, it depends on what suit the end requirements well. But just a comparison of fload and mload for empty tables and no indices, fload is a bit faster(though both are close), since it does not log extra information compared to mload. This is my understanding.
Cheers,
Raja
 

Raja K Thaw
My wiki: http://en.wikipedia.org/wiki/User:Kt_raj1
Street Children suffer not by their fault. We can help them if we want.

s@ir@m 35 posts Joined 05/13
18 Dec 2013

Hi Raja,
if i ahve a data like 
total 100 records
1. 50 records complete row duplicated.
2. 10 records data type is different .
3. 40 records unique and good data.
 
my question is 
Q : using fastload and multiload what are the error tables and log information ?
 
 
Regards
Ratnam

Raja_KT 1246 posts Joined 07/09
18 Dec 2013

Hi Ratnam,

We use mload and fload for reasons.

 

I hope this is for testing only :).Teradata is primarily for huge volume of data.

 

 

Mload:

 

LOGTABLE ---Identifies the table to be used to checkpoint information required for safe, automatic restart of Teradata MultiLoad when the client or Teradata Database system fails.

 

You have Work table,Error table,UV table,Target table: 

 

In dml LABEL, you can see mark/ignore duplicate. It is your choice.

mload has many tracking options to log.

 

For learning, check  the table description and values, since a real time experience is galaxy away from fantasy. Implement your cases and see.

 

FLOAD: 

 

Error table1 : Constraint violations,Conversion errors,Unavailable AMP conditions

 

Error table2: Unique PI violation

 

Check the fields and try to analyze for both tables.

 

You can go thru the materials of mload and fastload and you will see they have loads of information.

(My suggestion is create sample data....implement them with all your ifs and buts)

 

Cheers,

Raja

Raja K Thaw
My wiki: http://en.wikipedia.org/wiki/User:Kt_raj1
Street Children suffer not by their fault. We can help them if we want.

s@ir@m 35 posts Joined 05/13
19 Dec 2013

thku Raja,
Q : compleate row duplicates 50 records in my file  using fastload after executing script log information has given 49 record as duplicated count and 1 record will go to UV  error table why ? can you give me the clarifcation .
 
ratnam

Raja_KT 1246 posts Joined 07/09
20 Dec 2013

Check and verify your data source.
Also check and verify :
SELECT  * FROM errortname1
and SELECT * FROM errortname2.
Cheers,
Raja

Raja K Thaw
My wiki: http://en.wikipedia.org/wiki/User:Kt_raj1
Street Children suffer not by their fault. We can help them if we want.

SmarakDas 51 posts Joined 02/12
16 May 2014

Hello All,
 
This is an interesting question concerning FastLoad & MultiLoad. I also created a similar forum post on a broad topic. Would appreciate if anyone can provide their feedback on it.
 
http://forums.teradata.com/forum/tools/difference-between-load-export-utilities
 
Thanks In Advance,
Smarak

You must sign in to leave a comment.