Breniser

when create new user database without rights, new user can still see schemas/tables etc. can't access them, still sees them. revoke these privileges, not know how. this how created user: create user wouter_test password 'wouter_test' nosuperuser nocreatedb nocreaterole nocreateuser inherit; based on post thought may have rights users have public schema , information public schema contains: https://dba.stackexchange.com/questions/98892/minimal-grants-for-readonly-single-table-access-on-postgresql based on wikisite: https://wiki.postgresql.org/wiki/shared_database_hosting used command revoke on schema public wouter_test; it did not work. following did not seem work either (to prevent user seeing , accessing database called klm ) revoke connect on database klm wouter_test; but still user, in pgadmin, can see databases, schemas , tables (including klm ). what doing wrong? you can revoke ... public forbid users use object, cannot keep user se...

as latest version of spark (2.1.0) lists unsupported major functionality of hive like: tables buckets: bucket hash partitioning within hive table partition. spark sql doesn’t support buckets yet. it means hive tables have bucketed column structure won't loaded dataframe processed properly. there workaround achieve such functionality via jdbc , temp tables , etc. main question how achieve full control using spark sql on tables materialized deltas ?

i working on odata service sap fiori application. when try code bapi inserting notifications database, doesn't work & don't know why. colleague told me it's because of missing mandatory parameters filled them , no result. here's code : method avisset_create_entity. data: ls_data type zcl_zpm_avis_mpc=>ts_avis. data: l_notif type bapi2080_nothdre-notif_no, l_notif_type type bapi2080-notif_type, l_notif_hd_exp type bapi2080_nothdre, l_notif_header type bapi2080_nothdri, i_return type table of bapiret2. io_data_provider->read_entry_data( importing es_data = ls_data ). if ls_data not initial. l_notif_type = 's3'. l_notif_header-equipment = ls_data-equnr. l_notif_header-reportedby = sy-uname. l_notif_header-short_text = ls_data-qmtxt. * l_notif_header-notif_date = ls_data-qmdat. l_notif_header-code_group = 'maindiag'. l_notif_header-coding = 'desi'. * l_notif_header-cod...

i new webpacks. started working rails 5.1.0beta1 version comes webpack support. deploying on heroku fails following error during asset compilation remote: sh: 1: ./node_modules/webpack/bin/webpack.js: not found remote: rake aborted! remote: json::parsererror: 743: unexpected token @ '' can please suggest how fix it. have tried doing this: heroku buildpacks:set heroku/nodejs --index 1 heroku buildpacks:set heroku/ruby --index 2

what want is, 2 different cache implementations (let's redis , ehcache) on 1 method. meaning @cacheable method should cache both redis , ehcache. is possible? option 1: stack caches. configure local cache in spring. wire in distributed cache via cacheloader/cachewriter. consistency needs evaluated. e.g. if update goes distributed cache, how invalidate local caches? not trivial. maybe easy , not needed data, maybe near impossible. option 2: go distributed cache provides called near cache . combination want yourself, combined in product. know hazelcast , infinispan offer near cache. however, mileage may vary regarding consistency , resilience. know hazelcast enhanced near cache, consistent. interesting use case , common problem. further thoughts , discussion highly appreciated.

i using sqlite-net library xamain , have question. call better or faster? var item = sqliteconnection.table<user>().firstordefault(e => e.id == myobject.id); or var item = sqliteconnection.table<user>().where(e => e.id == myobject.id).firstordefault(); either fine. they both run lazily - if source list has million items, tenth item matches both iterate 10 items source. performance should identical , difference totally insignificant. but i started getting "the wait operation timed out." using where(query).firstordefault() on large dataset. when changed firstordefault(query) stopped seeing it the difference when compare 2 approaches working out 1ms on 150,000 items (not 1ms each item, it's 1ms 150,000). based on 500 x 1,000 items query that's 0.003ms per day. it's not significant but prefer firstordefault(query)

i'm working on xamarin forms project in need able user's phone number device. figured out not possible on ios , windows phone, still want android, that's why i've found on internet piece of code should working on android. i'm using mvvm pattern. in viewmodel, try phone number way : string phonenumber = dependencyservice.get<iphonenumber>().getphonenumber(); which calls method class : public class phonenumber : iphonenumber { public string getphonenumber() { telephonymanager mtelephonymgr; mtelephonymgr = (telephonymanager)application.context.getsystemservice(context.telephonyservice); return mtelephonymgr.line1number; } } when run app on android emulator visual studio, works, phone number made of 11 digits. when i'm on real device, method returns empty string (""). i've put read_phone_state permission in androidmanifest.xml file. oh also, of course,...