Jump to content

AdamS

Members
  • Content Count

    598
  • Joined

  • Last visited

  • Days Won

    22

AdamS last won the day on March 27

AdamS had the most liked content!

Community Reputation

30 Excellent

About AdamS

  • Rank
    Advanced Member

Contact Methods

  • Website URL
    https://www2.deloitte.com/au/en/footerlinks/office-locator/western-australia-deloitte-australia-office-details/perth.html

Profile Information

  • Gender
    Not Telling
  • Location
    Perth

Recent Profile Visitors

936 profile views
  1. Neil it would be a simple matter to use Intella's inbuilt ability to detect foreign language documents, isolate the 5000 Spanish documents, then export only those documents into the load file.
  2. Chris the only time I've seen this before was when the indexing process was interrupted or failed to complete properly. I would try reindexing (if you haven't already). Also for urgent issues I'd always suggest starting a ticket or emailing support directly, as much as the team try to monitor and reply here I have noticed that from time to time they are not as active as I'd not they'd like to be. I suspect close to update release time they get caught up in the work and can't be here as much....
  3. Hi Jon, do you know if this is high on the priority list to fix, could I humbly suggest bumping it up if it's not? While the work around is simple enough for a couple of documents it's not really something I'd like to contemplate on your average discovery matter with thousands (tens of) documents. Added to that I feel safe in thinking just about any set of documents is going to have this issue. Is there a way for us to identify documents which may have landscape pages en masse with Intella currently?
  4. Corey do you use batching to assist your customers with the review process? I think you could completely exclude irrelevant items when creating the batches, not precisely what you are asking for I know but it's a workable solution, this method would also negate the need for the client to worry about the dedupe button.
  5. Okay Corey, I've found myself with some time on my hands so ran a test on some existing data which I had already used to create a Relativity load file. Below are the steps I took and the results achieved when importing the overlay back into the case. First step is to setup custom columns, I like to do this prior to indexing my source data. Basically add sources but do not index, then once the custom columns are setup go back and 'reindex' the data. You don't have to do this, you can index your data then add custom columns, however you will need to reindex after adding custom columns so you are better off time wise to index only once if you can set your workflow that way. 1. Custom Columns For what you are trying to achieve I would setup the following custom columns (the names are not important, those are just what I use). DOCID PARENT_DOCID ATTACH_DOCID BEGINGROUP ENDGROUP I set these up without mapping them to any specific fields in the Properties or RAW DATA tabs, so I just use the options like the pic below, this just creates a column without any data populated. They need to be kept blank in this instance so I can map them when I overlay the loadfile data. 2. Loadfile Creation After the data is indexed and the loadfile data selection is done it's time to create my loadfile, ensuring I use the same names as my custom columns for continuity, the pics below show the field selection options I chose for these fields when creating the load file: DOCID - Default field created when you setup the loadfile, depending on your naming conventions it will be something like PREFIX.000001.0000023 PARENT_DOCID - exactly what it sounds like, the same naming convention but reflecting the parent document ATTACH_DOCID - as you would expect, the child items BEGINGROUP - This is the first file in a group, which will be the parent ENDGROUP - The final file in a group, ie the last child item For the purpose of this discussion you can ignore the BATES numbering columns. The below picture is an excerpt of my load file showing the relationships are captured at the loadfile creation using the DOCID numbers. 3. Import .DAT Loadfile as Overlay Now go to File-->Import Loadfile Change the Import Operation to "Overlay", then use the "..." button to navigate to your .DAT file from the loadfile you just created and click 'next'. You should now have something like the screen below Now we need to map the columns from the loadfile to the custom columns you created. Click 'Next' then begin the mapping exercise You don't have to map all the fields, just the ones you need, in the right hand box under 'Intella Columns' locate the custom fields you created, highlight then click the left blue arrow to move that across to match the corresponding load file field you want to map it to. Once you have all the desire fields mapped click 'import'. IMPORTANT - it's very important the you have the overlay options (top left) field correctly setup otherwise the import will not work. I have use the MD5 hash as this is part of my loadfile, however the ITEM_ID would be the logical choice provided this is part of your loadfile. Screenshot below shows my new fields exposed buy unpopulated at the beginning of the overlay import (the column with numbers is the Item ID field). This overlay import was for 26k items and took about 15 minutes. The screenshot below shows after the process has completed and you can see the new fields have been populated. Now to address your specific needs This is now achievable simply by sorting by the DOCID field 1. Sorting by DOCID achieves this 2. The BEGINGROUP is what you want here I think, this is the parent file and first file in a family 3. I'm not sure I understand the goal here. The very nature of hashes is to be unique so it would go against all forensic procedures to have different files with the same hash. What sort of custom deduping are you wanting. Perhaps this can be achieved now with the custom fields. This is another one I'm not sure that I'm understanding correctly. If you mean if you have 2 emails which both contain exactly the same children items? If that's the case a normal dedupe would take care of that now as all items would have duplicates. If you mean you have 2 identical parent emails but then have slightly different children and you want to only keep one of those parents and the resulting children I believe you can accomplish that now by using top level deduplication then tagging only children of the deduped top level items. Perhaps you could elaborate on this point a bit as I'm not sure what your goal is here. Intella has a fairly strong customisation ability for loadfile creation already, and in fact I found it's just as extensive (if not better) than the other common tool used for processing data. I've found that eDiscovery can vary greatly from client to client as to what they would like to see, however you are correct that there are many data points which are standard across load files, even if the end client doesn't see them on the review side of things. I hope I have at least answered some of your questions
  6. Corey I suspect you can accomplish some (if not all) of this currently by using custom columns and then creating an overlay to import back in to the case. I currently have quite a few custom columns that I add to a case before I index any data to avoid the need for reindexing. The parent/child relationship is created when preparing Loadfiles, so I suspect there a way you can leverage this back into the case with an overlay. Sorry I don't have the time at the moment to run any tests and provide you something definitive, but I suspect Jon won't be too far behind in providing something more substantial.
  7. While you could spend much time testing and trying beta drivers or other tricks, I think ultimately the best solution here would be to work through with the Intella Dev guys to find a permanent fix. I've never seen anything remotely like what you are experiencing, but it would seem that from time to time Nvidia and Intella don't play nice together. If you are desperate for a short term fix I would suggest possibly going out and grabbing a cheap video card that with an ATI chip, remove Nvidia from the equation until a suitable fix can be found.....or possibly Beta drivers..
  8. I didn't realise this either, and it's something I've been wanting for ages haha
  9. The parent_tab/child_tag syntax will assist in the short term, however to avoid the need having the ability to simply tick a box and type a 'Parent' name as part of the AutoTag process would be ideal. The changes I'm thinking of are giving the ability to apply parameters to keyword lists for searching and auto tagging. So for comparison, if I want to search across the subject lines ONLY of all emails I do the following steps: Highlight emails in the type-->communications facet, right click and select 'include' select the 'options' button next to the free text search field (top left) and untick all options except 'subject' enter search term in the free text search field and press enter results display matches across emails only in the subject field I can then highlight and tag those results before moving on to the next search term Using that as a basis for what I'm looking for imagine how to make that possible using a keyword list so we can avoid typing in hundreds of individual search terms. When we add a keyword list and then select 'AutoTag' the only option we can change is the tagging rules (item only, including child items or including all family tree items). There will be many uses I think where having the ability to apply the following filters to keyword lists AND have them auto tag would be great: Dedupe/ignore irrelevant Specify fields to search across (subject, text/body, email addresses etc) One tag only per keyword per item/document First occurrence applies within single item (single item has multiple keywords which are responsive, only the first responsive keyword tag is recorded) Highest number of hits on single document (single item has multiple keywords responsive, keyword with the highest number of occurrences is recorded) Just a few thoughts but I'm sure there are other ways we could give some granular control over keyword list auto tagging.
  10. The autotag feature when undertaking keyword list searches is something I use quite a bit, but it would be great to see some further control over how those tags are applied. Lets say I already have a dozen or so tags applied, and a structure built up, if I import a new keyword list and want to auto tag those new tags are going to be interspersed between my existing tags and basically make a mess and trigger my neatness OCD bug To whit, currently we can set the tagging preference only (ie tag the selected item only, tag it's children as well etc..) I would be great to simply have a check box and be able to put that list of new tags into it's own nested parent tag, for example all these new keyword tags would be under a parent tag 'KWS 2' or what ever we want to call it. That way once the search and subsequent tag operation has completed it is very easy to work with that new set of tags. Secondly we can't influence what those new keywords are searching across, it would be another good addition if we could apply exceptions/inclusions or the options filter to the autotag/search process to limit the false positives and the cleanup that we have to undertake after the operation.
  11. This might seem a little trivial considering all we need to do is copy the .xml file to the correct folder in the Appdata location, however once we do that we need to reload the case for it to be visible. If an 'import export template/profile' button existed then this may negate the need to reload the case...?
  12. Expanding on the above query. I'm attempting to have an all encompassing field like the Primary Date, but them manipulating the way the data displays so instead of the full date and time it will just display the UTC offset, say 1000 or AEST for Sydney. In my testing I tried creating a custom field which looked for the email sent record ( PR_CLIENT_SUBMIT_TIME ) field in RAW DATA and also in the Headers field, then changed the date/time format using z(Z) or just z or Z and a few other variations. After re-indexing I cannot get this field to populate with any data at all. I have tried about 10 different date/time formats and even reverted back to the default. I have tried being specific to emails only or 'ANY' in the setup options. I have all the other custom fields setup and working correctly (email importance, sensitivity, read receipt request etc), this is the only one that is causing me problems. Any advice appreciated.
  13. I'm running some load file testing attempting to duplicate load files created with other software and I have a question. Essentially what I'd like to do is have the ability to duplicate the Primary Date column, but have the date format display in a couple of different ways. My thought was to create a custom field, then duplicate the fields which are listed in the Preferences section for Primary Date, however there is not enough information in the Preferences section for me to duplicate the Primary Date field. Is there a file somewhere that shows precisely which fields within Raw Data (or elsewhere) that are being parsed for the Primary Date field?
×
×
  • Create New...