mongodb - Detecting concurrent data modification of document between read and write -

i'm interested in scenario document fetched database, computations run based on external conditions, 1 of fields of document gets updated , document gets saved, in system might have concurrent threads accessing db.

to make easier understand, here's simplistic example. suppose have following document:

{    ...    items_average: 1234,    last_10_items: [10,2187,2133, ...]    ... } 

suppose new item (x) comes in, 5 things need done:

1. read document db
2. remove first (oldest) item in last_10_items
3. add x end of array
4. re-compute average* , save in items_average.
5. write document db

* note: average computation chosen simple example, question should take account more complex operations based on data existing in document , on new data (i.e. not solvable $inc operator)

this easy implement in single-threaded system, in concurrent system, if 2 threads follow above steps, inconsistencies might occur since both update last_10_items , items_average values without considering and/or overwriting concurrent changes.

so, question how can such scenario handled? there way check or react-upon fact underlying document changed between steps 1 , 5? there such thing watch redis or 'concurrent modification error' relational dbs?

thanks

in database system,it uses memory inspection , roll scheme similar transactional memory.

briefly speaking, monitors share memory parts specified , compare , swap or load , link or test , set. therefore,if memory content changed during transaction,it abort , try again until there no conflict operation shared memory.

for example,gcc implements following:

https://gcc.gnu.org/onlinedocs/gcc-4.1.2/gcc/atomic-builtins.html

type __sync_lock_test_and_set (type *ptr, type value, ...) type __sync_val_compare_and_swap (type *ptr, type oldval type newval, ...) 

for more info transactional memory, http://en.wikipedia.org/wiki/software_transactional_memory