How to merge duplicate nodes by label - duplicates

Gephi has an option to merge duplicate nodes. It asks which column I want to use to identify duplicates.
However, the Label column does not identify any duplicates despite there being some (e.g. it doesn't recognize two nodes labeled "Jim" as duplicates).
How to merge duplicate nodes by label? I have over 100,000 nodes.

Related

SSIS:I need to apply fuzzy lookup on multiple table columns

I need to apply fuzzy lookup on multiple table columns. for example - I have table A which contains 4 columns(50% matched data ) which look 4 different tables which contain 100% matched data. I want to apply fuzzy lookup on 4 different data sets which match data from different 4 tables and give me correct data for table A. How can I do this.
In the edit querys go to Merge Querys > Merge As New and check the Use fuzzy matching to compare the merge in the pane (You also have some fuzzy merge options here for example the match percentage) and hit OK.
If you have more tables to match with, just repeat the first step again on the newly created table.
You can also pass a transformation table where you cen specify some matching criterias.

Both inputs of the transformation must contain at least one sorted column, and those columns must have matching metadata ssis

I have created an SSIS package and I used Merge Join to join a Dimension with the result of another Merge Join and I got the following error:
Both inputs of the transformation must contain at least one sorted column, and those columns must have matching metadata ssis
I have found that the issue is related to the data type of the two sorted columns, I just made a conversion to make both of them "INT" and everything is going fine.
The message is pretty clear. SSIS merge operations required that the data to be compared is sorted so comparisons are faster.
Make sure that you are retrieving ordered data from your database using the ORDER BY clause (if on SQL), and mark the columns with their corresponding order at the property IsSorted.
If you can't have the data ordered at the source, you can add a Sort operation in SSIS which will sort the merging columns (before the actual merge). You will have to do this on both flows before the merge. Please be adviced that using this componene will block the data flow until all rows are sorted.
The Merge error message will go away once you join both data flows with sorted columns.

In SSIS, how can one simply ignore records that Lookup identifies as not a match?

In my current SSIS data flow task, I feed my data flow into a Lookup tool. Matches are inserted into one table and non-matches are inserted into another table.
I did it this way because this is what I was able to learn from the available tutorials at the time.
However, it seems wasteful because I don't really want the non-matched records at all. Is there a way to tell SSIS to discard the non-matched records entirely rather than store them in a table?
The lookup dialog doesn't appear to give me an option for "ignore non-matches."
Is there some way to achieve this desired behavior?
If lookup = match, insert matched records into table (as currently done)
If lookup not a match, ignore (or discard) non-matched records
Leave Redirect rows to no match output as you currently have specified.
Select the "non-matched" branch and delete the destination.
Done.
Really, that's it. The rows will still be in the memory buffers of your data flow but they won't carry to the Match destination as they'll be logically segmented.
Personally, I have a Row Count wired up so I can count the original Rows, the matched rows and the unmatched rows. It helps me audit how the package has performed over time but there's nothing wrong with not using an output stream from a component.

SQL : should I insert another column or parse every single row

Assume millions of lines of traffic data in SQL format.
From the column URL and for each row of given range, I want to get a substring text that matches the target tag.
For example, from the column URL, I have the following texts:
Column: `URL`
Row 1: http://www.google.com/abcdeft?&QQ=123&AA=america&YY=111
Row 2: http://www.google.com/abcdeft?&QQ=123&AA=asia&YY=111
Row 3: http://www.google.com/abcdeft?&QQ=123&AA=africa&YY=111
Row 4: http://www.google.com/abcdeft?&QQ=123&AA=south&YY=111
Row 5: http://www.google.com/abcdeft?&QQ=123&AA=south&YY=111
Row 6: http://www.google.com/abcdeft?&QQ=123&AA=&YY=111
Row 7: http://www.google.com/abcdeft?&QQ=123
...
Row 99999999: http://www.google.com/abcdeft?&QQ=123&AA=ddd&YY=111
Data keep being loaded with lots of updates. So performance does matter. My goal is to:
Identify each row with its unique key-tag &AA=. Basically I need to get the string in the tag &AA= from every single row. For example, I want africa from ~~&AA=africa&~~. None if there is no &AA= but still need to read every single row.
Identify duplicate rows that contain the same tag in &AA=. e.g. row 4 and 5 are duplicates because they have same AA tags of south.
Question: which would be the best way for future data processing?
Option 1. Without URL column
Read every single row in URL column
Parse each row for the tag &AA= using urlparse library
Need a separate script to find duplicate rows with the same AA tag. e.g. using Python, I need to make a list of all items(all tags) and find the duplicate items in the list.
Need a separate query to find the rows that contain duplicate tags. e.g. query the rows that contain the duplicate items in the column URL
Creating separate column specifically for this task seems relatively doable.
Option 2. Insert another new column AA for tag &AA= and start filling out the new column when updating traffic data.
In this way:
No need to Read the column URL
No need to Parse the text in URL to get the tag &AA=
No need to Find duplicate items from one query
- No need to etrieve rows with duplicate items from another query
In this way, we can easily:
Get &AA= data just selecting the column AA
SELECT duplicate rows using COUNT function in SQL
Which one would perform better?
If you can stand the extra space cost of having an additional column then that would be the optimal approach. If there are a lot of duplicates of AA you might consider putting that in another table and then joining to it for queries. That would cut down on the space cost and still give you all the flexiblity. it would make it even easier (faster to query) if you were querying on an ID instead of the textual value of AA.

Remove duplicates in SSIS Data Flow

I am working on an SSIS data flow task.
The source table is from old database which is denormalized.
The destination table is normalized.
SSIS fails because the data transfer is not possible because of duplicates (duplicates in primary key column).
It would be good if the SSIS can checks the destination for availability of current record (by checking the key) and if it exists , it can ignore pushing it. Then it can continue with the next record.
Is there a way to handle this scenario?
Assuming your destination table is a subset of your source table, you should be able to use the Sort Transformation to pull in only the columns you need for your destination table, and then check the "Remove rows with duplicate sort values" to basically give you a distinct list of records based on the columns you selected.
Then, simply route the results of the sort to your destination, and you should be good to go.