Freeze row headers/initial column(s) to make it easier to tell where you are in a sparsely populated grid

GoogleCodeExporter commented 8 years ago

*****
If you don't know how to do something in Google Refine, please ask on the 
mailing list. Also use the mailing list for discussions and comments.  
http://groups.google.com/group/google-refine/
Only file an issue here if you discover a bug or want to request a new
feature.  Thank you.
*****

>>What steps will reproduce the problem?
Step 1. Interact with a large data set with `many` rows (>5000) and `many` 
columns (~80). One column should be titled "other_type_6" or "OtherType6" 
depending on which way you swing (optional sub-step)

2. Most records should be empty, noticeable for their emptiness and not 
containing very much stuff, with a key value here and a loose comment there and 
just a few other bits scattered among the columns, no big deal.  If your data 
set is well-populated with distinct and recognizable data in an intuitive 
structure then this is not going to be a very disorienting step for you.  So 
you should really try to use-slash-imagine we are working with a very porous 
data set. Russian expense reports, Italian time sheets, etc.

3. Imagine your task is to "clean" the records and it requires you to scan 
across each row to notice where the wrong bits have landed in the wrong 
columns. There are also relationships among some very not close to each other 
records.  In certain cases the nearest exit may be located behind you. 

>>What is the expected output? 
I expect to have the option to freeze either one header row, one header column 
or both at the same time. Without the same level of expectation, I would also 
be delighted to find that I could split the screen horizontally or vertically 
and experience some synchronized scrolling as I navigated my very expansive 
grid!  VisiCalc circa 1979 spoiled me forever with its edge-cutting 8-bit 
user-experience.  

>>What do you see instead?
I see a naked grid without a comforting frames of context and way-finding.  I 
am lost, confused and unproductive.  (Yes, I know)  I see generalized 
non-specific ennui in the eyes of almost everyone (unrelated.)  I  see myself 
scrolling laboriously up/down left/right vainly trying to stay oriented, 
deserted in a vast and formless expanse of white and grey quadrille. It feels 
like I am trying to circumnavigate the globe with an etch-a-sketch and peyote.  
I just want to check am I and the right cell, just let me check one more time, 
then I'll have it. Realy should double check if this is  izek_number, no one 
really uses that.  I like to confirm that I'm living in the right cell for my 
next data scrubbing move.  I also see and this is quite clear, I see me 
spending more time on this task (and on this bug report, eh?... ok fine... 
feature request.. grumble grumble...1979...grumble grumble 8-bit VisiCalc 
grumble..) than ought strictly be necessary. 

>?>What version of Google Refine are you using?
2.0 r1836 !!

>>What operating system and browser are you using?
Windows 7 64-bit Ultimate SP-1 (c:\ver says 6.1.7601. thanks ver you rock. 
windows 7 version 6.1. duly noted.) 
Chrome 12.0.742.100 (which also rocks. word.

>>Is this problem specific to the type of browser you're using or it happens in 
all the browsers you tried?
This problem is definitely not browser specific.  It is broadly general and 
indifferent to all the browsers I tried.

oh last step 4. Try to rearrange the data, yo.  Por ejemplo, move the value 
from "type_other_6" to "first_future_payment_date."  Move the info from eight 
"counterparty" columns over into the 6 partially-corresponding "broker_dealer" 
columns because some schmo clearly doesn't know the difference between a 
counter-party and a broker-dealer.  Break up some of the notes in "description" 
to select the right value for "asset_class" and "physical_commodity".  Imagine 
that these columns are non-adjacent, even highly to very-extremely 
non-adjacently located, dig?  At this point I want to rearrange these 19 
columns to make the task just a little bit easier. I move the columns using 
move left, move right, unit or extent. To make things easier on myself.  I 
start by choosing the last column, the one I care about least and want to see 
last, if at all. Choose that one first. Click down arrow in the column heading, 
edit column, move column begin. Find the next to last, the 2nd most least 
interesting column, for convenience, click down arrow in the column heading, 
edit column, move column begin, find the third most esoteric data field and 
click down arrow in the column heading, edit column, move column begin.  
Nicely.  Do that for the next 16 data columns in order of increasing abstruse 
relevance and try to think of a nice logical order for these columns, but 
backwards.  And then click-mousety on each one of them, in reverse succession, 
three times.  Ok, great.  I changed my mind about one of them. It is all good. 
all of it.  Settlement_date better should go four parking spaces over to the 
right nestled between account_exec and the izek_number?  NP!(LOL,faces) I 
simply click down arrow in the column heading, edit column, move column right, 
click down arrow in the column heading, edit column, move column right, click 
down arrow in the column heading, edit column, move column right, click down 
arrow in the column heading, edit column, move column right.  Perfecto.  When 
the browser freezes in the middle of this process the error says a script has 
used way [sic] too much memory. Let's just shut that one down and launch a new 
one, K.  Shutting that down, just take me a sec. It only happens every three to 
fine minutes so its fine, all. good.  Awesome! this as such a cool tool, I 
really love it. It is going to be so great when it is finished. A-maze-ing.  I 
can not wait. literally. can't. wait.  I really dig it.  Like you, I am a 
*huuge* fan of regular expressions, just. like. you.  But I miss not having a 
big outer NOT token for some sweeet filtering.  But it is practically just as 
good to star all of the YES! matches and then "Facet" (so cool you verb'd 
another noun, uncle!) to the NotStarred records.  Elegant. Like a brand new 
Cadillac... Cimarron.  Boy howdy it feels good getting those columns arranged 
just so.  What was I doing again?  My job is to... oh!  nope, almost had it...  

semi-colon white space happy parenthesis  xxo

Please provide any additional information below.
I like you.  You are fun.  Is that a new hoodie?   Thanks. you too.

Original issue reported on code.google.com by aa...@habilisdesign.com on 18 Jun 2011 at 7:07

GoogleCodeExporter commented 8 years ago

TL;DR but looks like same as 
http://code.google.com/p/google-refine/issues/detail?id=38

Original comment by paulm%pa...@gtempaccount.com on 18 Jun 2011 at 7:19

GoogleCodeExporter commented 8 years ago

The column headers now freeze vertically (and scroll horizontally). But the row 
headers don't freeze. Is this good enough?

Original comment by dfhu...@gmail.com on 27 Sep 2011 at 9:39

GoogleCodeExporter commented 8 years ago

Marked as Post 2.5 for making row headers freeze as well.

Original comment by dfhu...@gmail.com on 27 Sep 2011 at 9:40

Added labels: Milestone-Post-2.5

GoogleCodeExporter commented 8 years ago

Since issue 38 covers column headers specifically we'll focus this on the 
remaining piece.

Original comment by tfmorris on 18 Sep 2012 at 5:58

Changed title: Freeze row headers/initial column(s) to make it easier to tell where you are in a sparsely populated grid
Changed state: Accepted
Added labels: Type-Enhancement
Removed labels: Milestone-Post-2.5, Type-Defect

ramyakalyan / google-refine

Freeze row headers/initial column(s) to make it easier to tell where you are in a sparsely populated grid #412