2

I have 4 different types of external logging data that can be opened in my desktop application. They are displayed as tables. Each of these have a few columns in common (e.g. timestamp), but most of the columns are completely different to the other types of logs.

Since it would be useful to view 'combined' versions of these tables in one single tab, I would like to implement this feature, but I'm struggling to find a usable solution.

Please comment if the use-case is unclear.

I've come up with 3 ideas, but I'm not satisfied with any of them, to be honest. Please feel free to suggest another option. Which way is the best to go about this problem?

First Idea: Naive Combination

Just add all unique columns (over 30 if all 4 types are opened in my case) into one table. While the visibility of columns can be toggled, it would still display many columns, because as I mentioned most of them are unique.

Pros

  • Easy to implement ;)

Cons

  • Way too many columns.
  • Contents of unique columns will be separated. E.g. all data of log type A might be on the left side of the table, all data of log type B on the right side.

Second Idea: Stacked Columns

Let's say I open two types of logs.

The table header will then consist of two rows: one for log type A, the other one for log type B. Each row contains the columns for this type, if they are unique. For non-unique columns two header rows would simply be merged into one.

The position of columns would be configurable for each header row. Resizing on the other hand would not.

Pros

  • Compact

Cons

  • Maybe hard for the user to understand at first.
  • One cell of a column can contain other type of information than the one below, for example. For this I could either slightly color the rows for identification of log type or add a textual/visual identifier of log type as the first column. (or both)

Third Idea: x Tables in one tab with a ghostly scrollbar

E.g. Put 2 tables side-by-side, synchronize scrollbars to be at the nearest same timestamp.

Pros

  • Probably easy to understand for the user

Cons

  • Self-moving scrollbars
  • Each table might only be able to display portion of columns
  • What if 4 types of logs are opened? Would a 2x2 grid be displayed?
  • In the end just an even worse version of Idea #1

Fourth idea: Your turn

2
  • can you post some visual mocks of your efforts so far? This will make it easier to parse through your attempts, and make the question more valuable to future viewers as well.
    – Mike M
    Jun 25, 2019 at 15:37
  • Why is it necessary to show the data in a table? Not knowing what the goal of the users is or how they would like to work with the data it is difficult to give an answer. But my thinking goes away from a table. You could use the common columns as table an maybe group some of the different attributes together in a cell, maybe even use expanders to show these informations only on demand etc.
    – BrunoH
    Jun 26, 2019 at 15:36

2 Answers 2

1

I assume rows from all 4 logs corresponds to each other w.r.t. time-stamp or common columns for that matter.

Actually, feasible design depends on - 1) How user wants to use this data. What is user's goal w.r.t. all 4 logs 2) How big screen users have 3) Max number of columns in all 4 log tables 4) Volume of data in all 4 logs

e.g. If users are taking common columns as reference and find corresponding data in all 4 logs then adding all columns in single table (keeping common ones at left and selection for unique columns on right side) makes sense. This would look same as few frozen columns in excel sheet.

At the same time you want to see multiple unique columns and find "too many columns" even after selecting columns you want to see, it seems contradictory. May be I have not understood what users would actually do with the screen....

Stacked columns seems confusing.

Experimenting with stacked columns seems feasible.

enter image description here

Option1: Select any on tab corresponding to each log for unique columns. Columns can be toggled on/off for view

enter image description here

Option2: Select unique columns from multiple logs

enter image description here

3
  • Thank you for your answer. Regarding your questions: 1) The primary goal is to be able to see, analyze & filter the events of all log types 2) All have at least 1920x1080 with ~22 inches 3) Disregarding common columns: Log1 = 6, Log2 = 9, Log3 = 16, Log4 = 19, combined (common columns subtracted): 44 4) That heavily depends on the settings active for the external logs, but generally Log1 and Log4 have the most volume. In your Option1, do I understand correctly that the user is only able to see the data of the currently selected unique columns (i.e. bottom row of the unique columns "grid"? Jun 26, 2019 at 10:23
  • Option1: Data from unique columns will always be visible. and for non unique columns, user will select any of the 4 logs. Once selected all columns from selected log will appear. Jun 28, 2019 at 8:58
  • Option 2: Data from unique columns will always be visible.For non unique columns, user can select one or more columns from various logs. Looking at your requirement that most data is is log 1 and 4. IMHO option 2 seems more feasible. Jun 28, 2019 at 9:00
1

Is it possible to interpret the data for the end user, meaning that he/she only sees the relevant conclusions, in stead of all the raw data?

This raw data could then be made available to the end user after e.g. after clicking some button.

Personally I do not enjoy having to process large tables of logging data, of which 95% is not relevant in most cases.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.