LAE

Note: This was originally posted by an inactive account. Content was preserved by moving under an admin account.

Originally posted by: what_to_do

Hi,

I am currently running a left join in both LavaStorm and Access yet both produce different result? This there a difference in the way LavaStorm and Access compare strings.

Find attach both the access and Lava storm graph....Am i doing something wrong?

Confused
Attachments:
Left Join.zip

Note: This was originally posted by an inactive account. Content was preserved by moving under an admin account.

Originally posted by: Tim Meagher

Hi,

Sorry, I don't have Access installed so can't do the comparison.
What are you expecting to see in the output - how many records?

You have selected to do a left-inner join (using the Join Type parameter "li") are you sure that this shouldn't be a left join (using the Join Type parameter "l")?

As a side note, rather than placing the sort & join nodes yourself, and needing to configure all of these, you should be able to simply use the join node of your choice.

For example, in the attached graph, I have included a "Join Left Inner" and a "Join Left" node for comparison.
I believe in your case you probably want to use the "Join Left" node - based on what it sounds like you want output.

Within these pre-configured join nodes you are able to specify whether or not sorting is to be performed by simply switching the "SortLeft" and "SortRight" parameters. For instance if you know that the input data is already sorted, there is no need to perform this step, so you could simply set these parameters both to False.

Let me know whether or not this helps.

Tim.
Attachments:
Join Left Inner - Join Left.brg

Note: This was originally posted by an inactive account. Content was preserved by moving under an admin account.

Originally posted by: what_to_do

Hi Tim,

Thank you for the response.

In Access I've set the Join properties as :

Include ALL records from �AddressId 1446� and only those records from �AddressId_1� where the join fields are equal.

I was expecting 3 rows on the output (output attached).
Attachments:
Left output result.txt

Note: This was originally posted by an inactive account. Content was preserved by moving under an admin account.

Originally posted by: Tim Meagher

I'm a bit confused ...

The Left join output you have provided seems to be exactly the same as the AffressId 1446.txt input that is being provided to the graph in the LAE.

Are there actually 2 data sources that are being joined as within the LAE graph, there is only one delimited file node which you are joining on itself.

From the sounds of your description however, you have another input source "AddressId_1" - where is this represented in the LAE graph?

Note: This was originally posted by an inactive account. Content was preserved by moving under an admin account.

Originally posted by: what_to_do

Hi Tim,

I was able to mimic an example in Lavastorm of what i require. But could you please have a look at the attached graph and advise why they behaved differently.

Thanks
Attachments:
Left Join v2.zip

Note: This was originally posted by an inactive account. Content was preserved by moving under an admin account.

Originally posted by: Tim Meagher

Hi,

Sorry.. I understand the issue now and am looking into this.
I will send more information when it is available.

Tim.

Note: This was originally posted by an inactive account. Content was preserved by moving under an admin account.

Originally posted by: what_to_do

Hi Tim,

Could you please advise me of the status of the unexpected Node behaviour and the timeline for a fix?


Thanks

Pete

Note: This was originally posted by an inactive account. Content was preserved by moving under an admin account.

Originally posted by: Tim Meagher

Hi,

This has identified an issue with the join node where, under unusual circumstances, a record will not be produced. Specifically, when the two rows being compared contained exactly the same data (including field indexes), one of the records was the last record in its input, the join keys were on different fields in the input records and the join type was not simply an inner join. It's actually a bit more in depth and obscure than this description, as it would only occur if there was a left (or left inner) join, and the right input reached it's last record first, or it was a right (or right inner) join and that left input reached it's last record first.

In any case, this fix will be included in the next maintenance release of the LAE.

In order to overcome this issue in the meantime, you could either:

• Temporarily add an additional field to one of the inputs � and not output this from the join, or remove it afterwards.
• Remove one field from one of the inputs prior to the join
• Or simply ensuring that the records have different field orders will also do the trick. For instance, if you open up the BRDViewer, locate the name of the last field, and then choose to emit the fields in a different order, the problem will not appear

Note: This was originally posted by an inactive account. Content was preserved by moving under an admin account.

Originally posted by: what_to_do

Thank you for the reply;

I’m not really sure if this is an unusual circumstance but not withstanding that could you please advise when the next maintenance release of LEA will be available.

In relation to overcoming this issue; I think I’ll need to hold off until the maintenance release is available.

Further to the above does MDA offer rewards for users who identify limitations in the LavaStorm software?

Note: This was originally posted by an inactive account. Content was preserved by moving under an admin account.

Originally posted by: rboccuzzi

This issue should be resolved in the next LAE maintenance release, which should be in Q3 of this year. As for rewards, we don't currently have anything in place, but appreciate the effort in following through to help us identify the issue.

We are considering enabling "User Reputation" functionality in this forum, which would allow the community to recognize active participants such as yourself, however.

Cheers
Rich