Mailing-List: contact user-help@cassandra.apache.org; run by ezmlm
Precedence: bulk
Reply-To: user@cassandra.apache.org
MIME-Version: 1.0
In-Reply-To: <CAKiMtbeMbtL1+X5hKQ3FAa_8+zD8S+u_crtnuP+QNxCAB3T4JA@mail.gmail.com>
References: <CAKiMtbcwcn-+zUm-yriCXWmJfREYi7pP8r4V2=z0Wrh-CxU6rQ@mail.gmail.com>
 <CABNXB2Ahi7uAPkasG79kCtH7USAFHeBKSc_5ZuXVgV3ENU3xRA@mail.gmail.com>
 <CAKiMtbe34PkFATB46J4GBgod1KeSRfXucmWsfMo2ZPptMnFPQQ@mail.gmail.com>
 <CAKiMtbdcQ=p3AUYOFQFugU9ourf+zQ0y5OLr_vrfgQsT6ZqVQw@mail.gmail.com>
 <CABNXB2CyHPFmXPUZyanWZu1e5nV+Za=Bjd-wrs0bOEyJp7trGQ@mail.gmail.com>
 <CAKiMtbek9jrv2yuQiAARbiLbrm_ivfqDTpPcnAw2V8n9fzyy=w@mail.gmail.com>
 <CAKiMtbeBjfVYhYbyH7VHP29ozRc6zeRcjUDLmPAuL2R_f_t=NA@mail.gmail.com>
 <15858fb7a97.11a46aeb8429388.7038892613575369355@winguzone.com>
 <CAKiMtbetQCi1BgJC+QQLErA_nDg1YeqOBUm_5Ai2jTJWtZciUA@mail.gmail.com>
 <CAKiMtbe4_Akj6E_9wXxgPeTrH_bGWKxvheSToQSiu9txTMGrJA@mail.gmail.com>
 <CAKiMtbcy3_NVZuka=W2_SGZ-s7uue5QEYXxk7TTmSLcZcLpntQ@mail.gmail.com>
 <CABckJc5Bo97k+pcR3oOXeChFS5uahpxcts05pgzBRkWi8wSoPA@mail.gmail.com> <CAKiMtbeMbtL1+X5hKQ3FAa_8+zD8S+u_crtnuP+QNxCAB3T4JA@mail.gmail.com>
From: Ali Akhtar <ali.rac200@gmail.com>
Date: Sat, 12 Nov 2016 20:44:10 +0500
Message-ID: <CAKiMtbe1iXTqiM_xQqtoEK1yGdG+XvD8_FqRSjQZVMYQ8DT-rQ@mail.gmail.com>
Subject: Re: Consistency when adding data to collections concurrently?
To: user@cassandra.apache.org
Content-Type: multipart/alternative; boundary=001a114422805fda6905411c7aa1
archived-at: Sat, 12 Nov 2016 15:44:24 -0000

--001a114422805fda6905411c7aa1
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: quoted-printable

Or it could even take Set<label> as the first bound var:

void addLabel(Set<Label> label, String id);


On Sat, Nov 12, 2016 at 8:41 PM, Ali Akhtar <ali.rac200@gmail.com> wrote:

> Andrew,
>
> I was thinking about setting up an accessor with that query and a bound
> variable ? which binds to the instance being added, e.g:
>
> @Query("UPDATE my_table SET labels =3D labels + ? WHERE id =3D ?")
> void addLabel(Label label, String id);
>
> Will that  work?
>
> On Sat, Nov 12, 2016 at 8:38 PM, Andrew Tolbert <
> andrew.tolbert@datastax.com> wrote:
>
>> You can do it in a SimpleStatement assuming you provide the CQL exactly
>> as you provided, but in a PreparedStatement it will not work because cql
>> prohibits provide bind values in collection literals.  For it to work yo=
u
>> could provide a List of UDT values in a bound prepared statement, i.e.:
>>
>>     UserType udtType =3D cluster.getMetadata().getKeysp
>> ace("k").getUserType("u");
>>     UDTValue value =3D udtType.newValue();
>>     value.setString(0, "data");
>>
>>     PreparedStatement p0 =3D session.prepare("UPDATE my_table SET labels=
 =3D
>> labels + ? where id =3D ?");
>>     BoundStatement b0 =3D p0.bind(*Lists.newArrayList(value)*, 0);
>>     session.execute(b0);
>>
>> Thanks,
>> Andy
>>
>> On Sat, Nov 12, 2016 at 9:02 AM, Ali Akhtar <ali.rac200@gmail.com> wrote=
:
>>
>>> Looks like the trick was to use [] around the udt value literal.
>>>
>>> Any way to do this using the java driver?
>>>
>>> On Sat, Nov 12, 2016 at 7:58 PM, Ali Akhtar <ali.rac200@gmail.com>
>>> wrote:
>>>
>>>> Changing the double quotes to single quotes gives:
>>>>
>>>> UPDATE my_table SET labels =3D labels + {id: 'foo'} where id =3D '';
>>>>
>>>> InvalidRequest: Error from server: code=3D2200 [Invalid query]
>>>> message=3D"Invalid user type literal for labels of type list<frozen<la=
bel>>"
>>>>
>>>>
>>>> On Sat, Nov 12, 2016 at 7:50 PM, Ali Akhtar <ali.rac200@gmail.com>
>>>> wrote:
>>>>
>>>>> The question is about appending to a set of frozen<udt> and how to do
>>>>> that while avoiding the race condition.
>>>>>
>>>>> If I run:
>>>>>
>>>>>  UPDATE my_table SET labels =3D labels + {id: "foo"} where id =3D 'xx=
';
>>>>>
>>>>> I get:
>>>>>
>>>>> SyntaxException: line 1:57 no viable alternative at input '}' (...=3D
>>>>> labels + {id: ["fo]o"}...)
>>>>>
>>>>> Here labels is set<frozen<label>>
>>>>>
>>>>> On Sat, Nov 12, 2016 at 7:40 PM, Vladimir Yudovin <
>>>>> vladyu@winguzone.com> wrote:
>>>>>
>>>>>> If I used consistency =3D ALL both when getting the record, and when
>>>>>> saving the record, will that avoid the race condition?
>>>>>> If I use consistency level =3D all, will that cause it to end up wit=
h
>>>>>> [1,2]?
>>>>>> No. Even if you have only one host it's possible that two threads
>>>>>> first both read data and than overwrite existing value one by one.
>>>>>>
>>>>>> The list is actually of a list<frozen<my_udt>> and not a text (I use=
d
>>>>>> text for simplification, apologies).
>>>>>> In that case, will updates still merge the list values instead of
>>>>>> overwriting them?
>>>>>> Do you mean UPDATE cql operation? Yes, it adds new values to list,
>>>>>> allowing duplicates.
>>>>>>
>>>>>> When setting a new value to a list, C* will do a read-delete-write
>>>>>> internally e.g. read the current list, remove all its value (by a ra=
nge
>>>>>> tombstone) and then write the new list.
>>>>>> As I mentioned duplicates are allowed in LIST, and as DOC says:
>>>>>>
>>>>>> These update operations are implemented internally without any
>>>>>> read-before-write. Appending and prepending a new element to the
>>>>>> list writes only the new element.
>>>>>>
>>>>>> Only when using index
>>>>>>
>>>>>> When you add an element at a particular position, Cassandra reads th=
e
>>>>>> entire list, and then writes only the updated element. Consequently,=
 adding
>>>>>> an element at a particular position results in greater latency than
>>>>>> appending or prefixing an element to a list.
>>>>>>
>>>>>>
>>>>>> Best regards, Vladimir Yudovin,
>>>>>>
>>>>>> *Winguzone <https://winguzone.com?from=3Dlist> - Hosted Cloud
>>>>>> CassandraLaunch your cluster in minutes.*
>>>>>>
>>>>>>
>>>>>> ---- On Sat, 12 Nov 2016 07:57:36 -0500*Ali Akhtar
>>>>>> <ali.rac200@gmail.com <ali.rac200@gmail.com>>* wrote ----
>>>>>>
>>>>>> The labels collection is of the type set<frozen<label>> , where labe=
l
>>>>>> is a udt containing: id, name, description , all text fields.
>>>>>>
>>>>>> On Sat, Nov 12, 2016 at 5:54 PM, Ali Akhtar <ali.rac200@gmail.com>
>>>>>> wrote:
>>>>>>
>>>>>> The problem isn't just the update / insert though, right? Don't
>>>>>> frozen entities get overwritten completely? So if I had [1] [2] bein=
g
>>>>>> written as updates, won't each update overwrite the set completely, =
so i'll
>>>>>> end up with either one of them instead of [1,2]?
>>>>>>
>>>>>> On Sat, Nov 12, 2016 at 5:50 PM, DuyHai Doan <doanduyhai@gmail.com>
>>>>>> wrote:
>>>>>>
>>>>>> Maybe you should use my Achilles mapper, which does generates UPDATE
>>>>>> statements on collections and not only INSERT
>>>>>> Le 12 nov. 2016 13:08, "Ali Akhtar" <ali.rac200@gmail.com> a =C3=A9c=
rit :
>>>>>>
>>>>>> I am using the Java Cassandra mapper for all of these cases, so my
>>>>>> code looks like this:
>>>>>>
>>>>>> Item myItem =3D myaccessor.get( itemId );
>>>>>> Mapper<Item> mapper =3D mappingManager.create( Item.class );
>>>>>>
>>>>>> myItem.labels.add( newLabel );
>>>>>> mapper.save( myItem );
>>>>>>
>>>>>> On Sat, Nov 12, 2016 at 5:06 PM, Ali Akhtar <ali.rac200@gmail.com>
>>>>>> wrote:
>>>>>>
>>>>>> Thanks DuyHai, I will switch to using a set.
>>>>>>
>>>>>> But I'm still not sure how to resolve the original question.
>>>>>>
>>>>>> - Original labels =3D []
>>>>>> - Request 1 arrives with label =3D 1, and request 2 arrives with lab=
el
>>>>>> =3D 2
>>>>>> - Updates are sent to c* with labels =3D [1] and labels =3D [2]
>>>>>> simultaneously.
>>>>>>
>>>>>> What will happen in the above case? Will it cause the labels to end
>>>>>> up as [1,2] (what I want) or either [1] or [2]?
>>>>>>
>>>>>> If I use consistency level =3D all, will that cause it to end up wit=
h
>>>>>> [1,2]?
>>>>>>
>>>>>> On Sat, Nov 12, 2016 at 4:59 PM, DuyHai Doan <doanduyhai@gmail.com>
>>>>>> wrote:
>>>>>>
>>>>>> Don't use list, use set instead. If you need ordering of insertion,
>>>>>> use a map<timeuuid,text> where timeuuid is generated by the client t=
o
>>>>>> guarantee insertion order
>>>>>>
>>>>>> When setting a new value to a list, C* will do a read-delete-write
>>>>>> internally e.g. read the current list, remove all its value (by a ra=
nge
>>>>>> tombstone) and then write the new list. Please note that prepend & a=
ppend
>>>>>> operations on list do not require this read-delete-write and thus pe=
rforms
>>>>>> slightly better
>>>>>>
>>>>>> On Sat, Nov 12, 2016 at 11:34 AM, Ali Akhtar <ali.rac200@gmail.com>
>>>>>> wrote:
>>>>>>
>>>>>> I have a table where each record contains a list<string> of labels.
>>>>>>
>>>>>> I have an endpoint which responds to new labels being added to a
>>>>>> record by the user.
>>>>>>
>>>>>> Consider the following scenario:
>>>>>>
>>>>>> - Record X, labels =3D []
>>>>>> - User selects 2 labels, clicks a button, and 2 http requests are
>>>>>> generated.
>>>>>> - The server receives request for Label 1 and Label 2 at the same
>>>>>> time.
>>>>>> - Both requests see the labels as empty, add 1 label to the
>>>>>> collection, and send it.
>>>>>> - Record state as label 1 request sees it: [1], as label 2 sees it:
>>>>>> [2]
>>>>>>
>>>>>> How will the above conflict be resolved? What can I do so I end up
>>>>>> with [1, 2] instead of either [1] or [2] after both requests have be=
en
>>>>>> processed?
>>>>>>
>>>>>>
>>>>>>
>>>>>
>>>>
>>>
>>
>

--001a114422805fda6905411c7aa1
Content-Type: text/html; charset=UTF-8
Content-Transfer-Encoding: quoted-printable

<div dir=3D"ltr">Or it could even take Set&lt;label&gt; as the first bound =
var:<div><br></div><div><div style=3D"font-size:12.8px">void addLabel(Set&l=
t;Label&gt; label, String id);</div></div><div><br></div></div><div class=
=3D"gmail_extra"><br><div class=3D"gmail_quote">On Sat, Nov 12, 2016 at 8:4=
1 PM, Ali Akhtar <span dir=3D"ltr">&lt;<a href=3D"mailto:ali.rac200@gmail.c=
om" target=3D"_blank">ali.rac200@gmail.com</a>&gt;</span> wrote:<br><blockq=
uote class=3D"gmail_quote" style=3D"margin:0 0 0 .8ex;border-left:1px #ccc =
solid;padding-left:1ex"><div dir=3D"ltr">Andrew,<div><br></div><div>I was t=
hinking about setting up an accessor with that query and a bound variable ?=
 which binds to the instance being added, e.g:</div><div><br></div><div>@Qu=
ery(&quot;UPDATE my_table SET labels =3D labels + ? WHERE id =3D ?&quot;)</=
div><div>void addLabel(Label label, String id);</div><div><br></div><div>Wi=
ll that =C2=A0work?</div></div><div class=3D"HOEnZb"><div class=3D"h5"><div=
 class=3D"gmail_extra"><br><div class=3D"gmail_quote">On Sat, Nov 12, 2016 =
at 8:38 PM, Andrew Tolbert <span dir=3D"ltr">&lt;<a href=3D"mailto:andrew.t=
olbert@datastax.com" target=3D"_blank">andrew.tolbert@datastax.com</a>&gt;<=
/span> wrote:<br><blockquote class=3D"gmail_quote" style=3D"margin:0 0 0 .8=
ex;border-left:1px #ccc solid;padding-left:1ex"><div dir=3D"ltr">You can do=
 it in a SimpleStatement assuming you provide the CQL exactly as you provid=
ed, but in a PreparedStatement it will not work because cql prohibits provi=
de bind values in collection literals.=C2=A0 For it to work you could provi=
de a List of UDT values in a bound prepared statement, i.e.:<div><br></div>=
<div><div><div><font face=3D"monospace, monospace">=C2=A0 =C2=A0 UserType u=
dtType =3D cluster.getMetadata().getKeysp<wbr>ace(&quot;k&quot;).getUserTyp=
e(&quot;u&quot;);</font></div><div><font face=3D"monospace, monospace">=C2=
=A0 =C2=A0 UDTValue value =3D udtType.newValue();</font></div><div><font fa=
ce=3D"monospace, monospace">=C2=A0 =C2=A0 value.setString(0, &quot;data&quo=
t;);</font></div><div><font face=3D"monospace, monospace"><br></font></div>=
<div><font face=3D"monospace, monospace">=C2=A0 =C2=A0 PreparedStatement p0=
 =3D session.prepare(&quot;UPDATE my_table SET labels =3D labels + ? where =
id =3D ?&quot;);</font></div><div><font face=3D"monospace, monospace">=C2=
=A0 =C2=A0 BoundStatement b0 =3D p0.bind(<b>Lists.newArrayList(val<wbr>ue)<=
/b>, 0);</font></div><div><font face=3D"monospace, monospace">=C2=A0 =C2=A0=
 session.execute(b0);</font></div></div><div><br></div></div><div>Thanks,</=
div><div>Andy</div></div><div class=3D"m_-7708036090386924106HOEnZb"><div c=
lass=3D"m_-7708036090386924106h5"><div class=3D"gmail_extra"><br><div class=
=3D"gmail_quote">On Sat, Nov 12, 2016 at 9:02 AM, Ali Akhtar <span dir=3D"l=
tr">&lt;<a href=3D"mailto:ali.rac200@gmail.com" target=3D"_blank">ali.rac20=
0@gmail.com</a>&gt;</span> wrote:<br><blockquote class=3D"gmail_quote" styl=
e=3D"margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div di=
r=3D"ltr">Looks like the trick was to use [] around the udt value literal.<=
div><br></div><div>Any way to do this using the java driver?</div></div><di=
v class=3D"m_-7708036090386924106m_2260361352536298725HOEnZb"><div class=3D=
"m_-7708036090386924106m_2260361352536298725h5"><div class=3D"gmail_extra">=
<br><div class=3D"gmail_quote">On Sat, Nov 12, 2016 at 7:58 PM, Ali Akhtar =
<span dir=3D"ltr">&lt;<a href=3D"mailto:ali.rac200@gmail.com" target=3D"_bl=
ank">ali.rac200@gmail.com</a>&gt;</span> wrote:<br><blockquote class=3D"gma=
il_quote" style=3D"margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-lef=
t:1ex"><div dir=3D"ltr">Changing the double quotes to single quotes gives:<=
div><br></div><div><div>UPDATE my_table SET labels =3D labels + {id: &#39;f=
oo&#39;} where id =3D &#39;&#39;;</div><div><br></div><div>InvalidRequest: =
Error from server: code=3D2200 [Invalid query] message=3D&quot;Invalid user=
 type literal for labels of type list&lt;frozen&lt;label&gt;&gt;&quot;</div=
></div><div><br></div></div><div class=3D"m_-7708036090386924106m_226036135=
2536298725m_9104954310666033738HOEnZb"><div class=3D"m_-7708036090386924106=
m_2260361352536298725m_9104954310666033738h5"><div class=3D"gmail_extra"><b=
r><div class=3D"gmail_quote">On Sat, Nov 12, 2016 at 7:50 PM, Ali Akhtar <s=
pan dir=3D"ltr">&lt;<a href=3D"mailto:ali.rac200@gmail.com" target=3D"_blan=
k">ali.rac200@gmail.com</a>&gt;</span> wrote:<br><blockquote class=3D"gmail=
_quote" style=3D"margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:=
1ex"><div dir=3D"ltr">The question is about appending to a set of frozen&lt=
;udt&gt; and how to do that while avoiding the race condition.<div><br></di=
v><div>If I run:</div><div><br></div><div><div>=C2=A0UPDATE my_table SET la=
bels =3D labels + {id: &quot;foo&quot;} where id =3D &#39;xx&#39;;</div><di=
v><br></div><div>I get:=C2=A0</div><div><br></div><div>SyntaxException: lin=
e 1:57 no viable alternative at input &#39;}&#39; (...=3D labels + {id: [&q=
uot;fo]o&quot;}...)</div></div><div><br></div><div>Here labels is set&lt;fr=
ozen&lt;label&gt;&gt;=C2=A0</div></div><div class=3D"m_-7708036090386924106=
m_2260361352536298725m_9104954310666033738m_7461973989115263948HOEnZb"><div=
 class=3D"m_-7708036090386924106m_2260361352536298725m_9104954310666033738m=
_7461973989115263948h5"><div class=3D"gmail_extra"><br><div class=3D"gmail_=
quote">On Sat, Nov 12, 2016 at 7:40 PM, Vladimir Yudovin <span dir=3D"ltr">=
&lt;<a href=3D"mailto:vladyu@winguzone.com" target=3D"_blank">vladyu@winguz=
one.com</a>&gt;</span> wrote:<br><blockquote class=3D"gmail_quote" style=3D=
"margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><u></u><div=
><div style=3D"font-size:10pt;font-family:Verdana,Arial,Helvetica,sans-seri=
f"><span><div><span style=3D"background-color:#ffffcc" class=3D"m_-77080360=
90386924106m_2260361352536298725m_9104954310666033738m_7461973989115263948m=
_8952484672542457719m_3105371754016739186highlight">If I used consistency =
=3D ALL both when getting the record, and when saving the record, will that=
 avoid the race condition?</span><br></div></span><span><div><span style=3D=
"background-color:#ffffcc" class=3D"m_-7708036090386924106m_226036135253629=
8725m_9104954310666033738m_7461973989115263948m_8952484672542457719m_310537=
1754016739186highlight">If I use consistency level =3D all, will that cause=
 it to end up with [1,2]?</span><br></div></span><div>No. Even if you have =
only one host it&#39;s possible that two threads first both read data and t=
han overwrite existing value one by one.<br></div><span><div><br></div><div=
><span style=3D"background-color:#ffffcc" class=3D"m_-7708036090386924106m_=
2260361352536298725m_9104954310666033738m_7461973989115263948m_895248467254=
2457719m_3105371754016739186highlight">The list is actually of a list&lt;fr=
ozen&lt;my_udt&gt;&gt; and not a text (I used text for simplification, apol=
ogies).<br></span></div><div><span style=3D"background-color:#ffffcc" class=
=3D"m_-7708036090386924106m_2260361352536298725m_9104954310666033738m_74619=
73989115263948m_8952484672542457719m_3105371754016739186highlight">In that =
case, will updates still merge the list values instead of overwriting them?=
<br></span></div></span><div>Do you mean UPDATE cql operation? Yes, it adds=
 new values to list, allowing duplicates.<br></div><span><div><br></div><di=
v><span style=3D"background-color:#ffffcc" class=3D"m_-7708036090386924106m=
_2260361352536298725m_9104954310666033738m_7461973989115263948m_89524846725=
42457719m_3105371754016739186highlight">When setting a new value to a list,=
 C* will do a read-delete-write internally e.g. read the current list, remo=
ve all its value (by a range tombstone) and then write the new list.<br></s=
pan></div></span><div>As I mentioned duplicates are allowed in LIST, and as=
 DOC says:<br></div><blockquote style=3D"border:1px solid rgb(204,204,204);=
padding:7px;background-color:rgb(245,245,245)"><div><div>These update opera=
tions are implemented internally<span style=3D"color:#ff0000" class=3D"m_-7=
708036090386924106m_2260361352536298725m_9104954310666033738m_7461973989115=
263948m_8952484672542457719m_3105371754016739186colour"> without any read-b=
efore-write</span>. Appending and prepending a new element to the list writ=
es only the new element.<br></div></div></blockquote><div>Only when using i=
ndex<br></div><blockquote style=3D"border:1px solid rgb(204,204,204);paddin=
g:7px;background-color:rgb(245,245,245)"><div><div>When you add an element =
at a particular position, Cassandra reads the entire list, and then writes =
only the updated element. Consequently, adding an element at a particular p=
osition results in greater latency than appending or prefixing an element t=
o a list.<br></div></div></blockquote><span><div><br></div><div id=3D"m_-77=
08036090386924106m_2260361352536298725m_9104954310666033738m_74619739891152=
63948m_8952484672542457719m_3105371754016739186"><div>Best regards, Vladimi=
r Yudovin, <br></div><div><i><a href=3D"https://winguzone.com?from=3Dlist" =
target=3D"_blank">Winguzone</a> - Hosted Cloud Cassandra<br>Launch your clu=
ster in minutes.</i></div></div><div><br></div></span><div class=3D"m_-7708=
036090386924106m_2260361352536298725m_9104954310666033738m_7461973989115263=
948m_8952484672542457719m_3105371754016739186zmail_extra"><div id=3D"m_-770=
8036090386924106m_2260361352536298725m_9104954310666033738m_746197398911526=
3948m_8952484672542457719m_31053717540167391861"><div><br></div><div>---- O=
n Sat, 12 Nov 2016 07:57:36 -0500<b>Ali Akhtar &lt;<a href=3D"mailto:ali.ra=
c200@gmail.com" target=3D"_blank">ali.rac200@gmail.com</a>&gt;</b> wrote --=
--<br></div></div><div><div class=3D"m_-7708036090386924106m_22603613525362=
98725m_9104954310666033738m_7461973989115263948m_8952484672542457719h5"><di=
v><br></div><blockquote style=3D"border-left:1px solid #cccccc;padding-left=
:6px;margin:0 0 0 5px"><div><div dir=3D"ltr">The labels collection is of th=
e type set&lt;frozen&lt;label&gt;&gt; , where label is a udt containing: id=
, name, description , all text fields.<br></div><div><div><br></div><div><d=
iv>On Sat, Nov 12, 2016 at 5:54 PM, Ali Akhtar <span>&lt;<a href=3D"mailto:=
ali.rac200@gmail.com" target=3D"_blank">ali.rac200@gmail.com</a>&gt;</span>=
 wrote:<br></div><div id=3D"m_-7708036090386924106m_2260361352536298725m_91=
04954310666033738m_7461973989115263948m_8952484672542457719m_31053717540167=
39186zmail_block"><br></div></div></div></div><blockquote style=3D"margin:0=
 0 0 0.8ex;border-left:1.0px rgb(204,204,204) solid;padding-left:1.0ex"><di=
v dir=3D"ltr">The problem isn&#39;t just the update / insert though, right?=
 Don&#39;t frozen entities get overwritten completely? So if I had [1] [2] =
being written as updates, won&#39;t each update overwrite the set completel=
y, so i&#39;ll end up with either one of them instead of [1,2]?<br></div><d=
iv><div><div><div><br></div><div><div>On Sat, Nov 12, 2016 at 5:50 PM, DuyH=
ai Doan <span>&lt;<a href=3D"mailto:doanduyhai@gmail.com" target=3D"_blank"=
>doanduyhai@gmail.com</a>&gt;</span> wrote:<br></div><blockquote style=3D"m=
argin:0 0 0 0.8ex;border-left:1.0px rgb(204,204,204) solid;padding-left:1.0=
ex"><p>Maybe you should use my Achilles mapper, which does generates UPDATE=
 statements on collections and not only INSERT<br></p><div><div><div><div>L=
e=C2=A012 nov. 2016 13:08, &quot;Ali Akhtar&quot; &lt;<a href=3D"mailto:ali=
.rac200@gmail.com" target=3D"_blank">ali.rac200@gmail.com</a>&gt; a =C3=A9c=
rit=C2=A0:<br></div><blockquote style=3D"margin:0 0 0 0.8ex;border-left:1.0=
px rgb(204,204,204) solid;padding-left:1.0ex"><div dir=3D"ltr"><div>I am us=
ing the Java Cassandra mapper for all of these cases, so my code looks like=
 this:<br></div><div><br></div><div>Item myItem =3D myaccessor.get( itemId =
);<br></div><div>Mapper&lt;Item&gt; mapper =3D mappingManager.create( Item.=
class );<br></div><div><br></div><div>myItem.labels.add( newLabel );<br></d=
iv><div>mapper.save( myItem );<br></div></div><div><div><br></div><div><div=
>On Sat, Nov 12, 2016 at 5:06 PM, Ali Akhtar <span>&lt;<a href=3D"mailto:al=
i.rac200@gmail.com" target=3D"_blank">ali.rac200@gmail.com</a>&gt;</span> w=
rote:<br></div><blockquote style=3D"margin:0 0 0 0.8ex;border-left:1.0px rg=
b(204,204,204) solid;padding-left:1.0ex"><div dir=3D"ltr"><div>Thanks=C2=A0=
DuyHai, I will switch to using a set.<br></div><div><br></div><div>But I=
9;m still not sure how to resolve the original question.<br></div><div><br>=
</div><div>- Original labels =3D []<br></div><div>- Request 1 arrives with =
label =3D 1, and request 2 arrives with label =3D 2<br></div><div>- Updates=
 are sent to c* with labels =3D [1] and labels =3D [2] simultaneously.<br><=
/div><div><br></div><div>What will happen in the above case? Will it cause =
the labels to end up as [1,2] (what I want) or either [1] or [2]?<br></div>=
<div><br></div><div>If I use consistency level =3D all, will that cause it =
to end up with [1,2]?<br></div></div><div><div><div><div><br></div><div><di=
v>On Sat, Nov 12, 2016 at 4:59 PM, DuyHai Doan <span>&lt;<a href=3D"mailto:=
doanduyhai@gmail.com" target=3D"_blank">doanduyhai@gmail.com</a>&gt;</span>=
 wrote:<br></div><blockquote style=3D"margin:0 0 0 0.8ex;border-left:1.0px =
rgb(204,204,204) solid;padding-left:1.0ex"><div dir=3D"ltr"><div>Don&#39;t =
use list, use set instead. If you need ordering of insertion, use a map&lt;=
timeuuid,text&gt; where timeuuid is generated by the client to guarantee in=
sertion order<br></div><div><br></div><div>When setting a new value to a li=
st, C* will do a read-delete-write internally e.g. read the current list, r=
emove all its value (by a range tombstone) and then write the new list. Ple=
ase note that prepend &amp; append operations on list do not require this r=
ead-delete-write and thus performs slightly better<br></div></div><div><div=
><div><div><br></div><div><div>On Sat, Nov 12, 2016 at 11:34 AM, Ali Akhtar=
 <span>&lt;<a href=3D"mailto:ali.rac200@gmail.com" target=3D"_blank">ali.ra=
c200@gmail.com</a>&gt;</span> wrote:<br></div><blockquote style=3D"margin:0=
 0 0 0.8ex;border-left:1.0px rgb(204,204,204) solid;padding-left:1.0ex"><di=
v dir=3D"ltr"><div>I have a table where each record contains a list&lt;stri=
ng&gt; of labels.<br></div><div><br></div><div>I have an endpoint which res=
ponds to new labels being added to a record by the user.<br></div><div><br>=
</div><div>Consider the following scenario:<br></div><div><br></div><div>- =
Record X, labels =3D []<br></div><div>- User selects 2 labels, clicks a but=
ton, and 2 http requests are generated.<br></div><div>- The server receives=
 request for Label 1 and Label 2 at the same time.<br></div><div>- Both req=
uests see the labels as empty, add 1 label to the collection, and send it.<=
br></div><div>- Record state as label 1 request sees it: [1], as label 2 se=
es it: [2]<br></div><div><br></div><div>How will the above conflict be reso=
lved? What can I do so I end up with [1, 2] instead of either [1] or [2] af=
ter both requests have been processed?<br></div></div></blockquote></div></=
div></div></div></blockquote></div></div></div></div></blockquote></div></d=
iv></blockquote></div></div></div></blockquote></div></div></div></div></bl=
ockquote></blockquote></div></div></div><div><br></div></div></div></blockq=
uote></div><br></div>
</div></div></blockquote></div><br></div>
</div></div></blockquote></div><br></div>
</div></div></blockquote></div><br></div>
</div></div></blockquote></div><br></div>
</div></div></blockquote></div><br></div>

--001a114422805fda6905411c7aa1--