shuffle related rows

sp4m

Hi,
I'm very new on Opensesame and therefore I might have a problem that is very easy to solve. But somehow I could not find a solution yet, so maybe someone could help me with this.
I need to randomize some items in a column, so every subject gets another sequence of items. The problem I have is, that this item should always be presented with another item.
I have a screenshot of the loop table:

I want to randomize the CS_closed, so that CS_closed for every subject is presented with another US_rating (US_rating consists of different images). CS_closed 1 (here: CS_Kanji1.bmp) should always be presented with the same US_rating for the respective subject. Plus, US_rating cannot be randomized because it is depending on the column valence and correct_response. I hope you can understand what I mean. Thanks!

Caro

sp4m

Edit: You can see that I tried to do it blockwise, but I need a full randomization, so this is not what I was looking for.

sp4m

Edit: Also tried it with coding:

But unfortunately Python always crashes.

Sylvain

Hello,

So if I understood you want to random one column while keeping the other constant?
If so a stategy could be:

• first you enter your initial matrix in a loop at the very beginning of your scrip (just to load it);
• then you extract these values, just random the first column in a new datamatrix, then copy this new randomized datamatrix in the subsequant loop.

Like this:

With the inline for randomizing and preparing the subsequant loop as:

from datamatrix import DataMatrix
import random

# create a list to stock the element from the column of the datamatrix (ie Load_Item_Loop)
# to randomize

list_column1_to_randomize = list()

# fill this list
for row in items['Load_Item_Loop'].dm.column1_name:
    list_column1_to_randomize.append(row)

# randomize the list
random.shuffle(list_column1_to_randomize)

# then create a new datamatrix that will be used for the subsequent loop where
# column1_name is randomized but the other column not randomized
# create the new datamatrix

new_dm = DataMatrix(length=len(list_column1_to_randomize))
# fill the first column with the randomized list
new_dm.column1_name = list_column1_to_randomize
# initialize and fill the two other columns
new_dm.column2_name = ''
new_dm.column3_name = ''

for i, row in enumerate(new_dm.column2_name):
    new_dm.column2_name[i] = items['Load_Item_Loop'].dm.column2_name[i]
    new_dm.column3_name[i] = items['Load_Item_Loop'].dm.column3_name[i]

# now copy the new_dm in new_loop (it will become the new values of this loop)
items['new_loop'].dm = new_dm

Find the script too maybe it will be more explicit.
Just be sure you change the two loop as sequential (I do not think it matters a lot but up to you)

Best,

Sylvain

sp4m

Hello Sylvain,

thank you for replying. I started the experiment you added and the first column is randomized, that's exactly what I want! But column 2 and 3 are presented sequentially (even if I change 'order' to 'random'). This should not be the case. Plus, the biggest problem I have is that the items in column 1 are presented two times, so they should be randomized but like this:
column 1, row 1 to 8 is randomized (e.g. 6, 4, 7, 3, 2, 1, 8, 5) and this order is attached to the items in column 2. Row 9 to 16 in column 1 and 2 should contain the same pairing as row 1 to 8. Does this make any sense? ':D

Caro

sp4m

And maybe you could tell me what is wrong with the code mentioned above? Because I do not understand why Python always crashes.

Sylvain

I do not really know why it crashes otherwise if I understood what you want you could try that:

from datamatrix import DataMatrix, operations
import random

# create a list to stock the element from the column of the datamatrix (ie Load_Item_Loop)
# to randomize

list_column1_to_randomize = list()

# fill this list
for row in items['Load_Item_Loop'].dm.column1_name:
    list_column1_to_randomize.append(row)

# randomize the list
random.shuffle(list_column1_to_randomize)

# then create a new datamatrix that will be used for the subsequent loop where
# column1_name is randomized but the other column not randomized
# create the new datamatrix

new_dm = DataMatrix(length=len(list_column1_to_randomize))
# fill the first column with the randomized list
new_dm.column1_name = list_column1_to_randomize
# initialize and fill the two other columns
new_dm.column2_name = ''
new_dm.column3_name = ''

for i, row in enumerate(new_dm.column2_name):
    new_dm.column2_name[i] = items['Load_Item_Loop'].dm.column2_name[i]
    new_dm.column3_name[i] = items['Load_Item_Loop'].dm.column3_name[i]

print(new_dm)

# now shuffle everything but keep row associated

final_dm = operations.shuffle(new_dm)

# repeat it n times and concatenate

repeat = 1
for r in range(repeat):
    dm_repeat = operations.shuffle(new_dm)
    final_dm = final_dm << dm_repeat



print(final_dm)


# now copy the new_dm in new_loop (it will become the new values of this loop)
items['new_loop'].dm = final_dm

Datamatrix function are on this page:
http://datamatrix.cogsci.nl/

And find the modified script.

Hope it will work this time.

sp4m

Hello Sylvain,

wow, thanks alot! This looks like it could be a very nice solution for my problem. So far I got it going by this code:

But I think your code looks much nicer and as if it could be used in a more general way. I will try it this way as well Thank you so much for your time and this nice piece of code!

Caro