Re: [racket-users] "table" data structure in Racket

['Paulo Matos' via Racket Users]

On 22/02/2019 04:05, travis.hinkel...@gmail.com wrote:
> After posing the question yesterday, I spent a little time poking around
> in the Github repository for Apache Arrow and came to the same
> conclusion, i.e., large project presumably facilitated by corporate backing.
> 

True, they are large projects... maybe even facilitated by corporate
backing but someone got them started and they probably got started
without corporate backing. I would say that you should do it if you are
motivated to do it for fun or if you need it. Start small, maybe create
just a very small subset of bindings that solve a specific problem.
Document it, publicize it, allow people to contribute. Maybe one day,
corporate backing will come.


Good luck!

-- 
Paulo Matos

-- 
You received this message because you are subscribed to the Google Groups 
"Racket Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to racket-users+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Re: [racket-users] "table" data structure in Racket

[travis . hinkelman]

After posing the question yesterday, I spent a little time poking around in 
the Github repository for Apache Arrow and came to the same conclusion, 
i.e., large project presumably facilitated by corporate backing.


On Thursday, February 21, 2019 at 4:28:55 PM UTC-8, Alex Harsanyi wrote:
>
>
>
> On Thursday, February 21, 2019 at 7:19:39 AM UTC+8, travis.h...@gmail.com 
> wrote:
>>
>> Hi All,
>>
>> I'm resurrecting this thread to ask if anyone in the Racket community has 
>> Apache Arrow on their radar. It seems like Apache Arrow might be gaining 
>> steam.
>>
>
>> Apache Arrow is a cross-language development platform for in-memory data. 
>>> It specifies a standardized language-independent columnar memory format for 
>>> flat and hierarchical data, organized for efficient analytic operations on 
>>> modern hardware. It also provides computational libraries and zero-copy 
>>> streaming messaging and interprocess communication. Languages currently 
>>> supported include C, C++, C#, Go, Java, JavaScript, MATLAB, Python, R, 
>>> Ruby, and Rust. [Source 
>>> 
>>> ]
>>
>>
>> I have no clue what it would take to make Racket a supported language. I 
>> also don't have a sense of what kind of demand the Racket community has for 
>> this sort of thing. Just curious if anyone is thinking about a Racket and 
>> Apache Arrow pairing.
>>
>>
> I had looked at Apache Arrow when it was first mentioned in this thread, 
> but the project is large and providing bindings seemed like a large task -- 
> the Java, Rust, JavaScript and Python bindings are large projects by 
> themselves.  I suspect Apache Arrow (along with numpy and other scientific 
> / data-processing packages) are backed by several large companies who 
> provide the development resources.
>
> Providing usable Racket bindings for Apache Arrow is not something one can 
> do as a hobby in a few weekends.  I ended up writing my own data frame 
> package, which is smaller, simpler and has the features and performance 
> levels that I need.
>
> Alex. 
>  
>
>> Thanks,
>>
>> Travis 
>>
>> On Wednesday, April 6, 2016 at 1:27:56 AM UTC-7, Konrad Hinsen wrote:
>>>
>>> On 05/04/2016 21:12, Asumu Takikawa wrote: 
>>>
>>> > I haven't built anything like that, but I was hoping that we could get 
>>> a GSoC 
>>> > student for it (that didn't pan out though obviously). The idea was to 
>>> use 
>>> > packages from Python/Julia/R as inspiration: 
>>> > 
>>> >http://pandas.pydata.org/pandas-docs/stable/index.html 
>>> >http://dataframesjl.readthedocs.org/en/latest/ 
>>> >https://github.com/Rdatatable/data.table/wiki 
>>>
>>> Also consider this one: 
>>>
>>> https://arrow.apache.org/ 
>>>
>>> Konrad. 
>>>
>>

-- 
You received this message because you are subscribed to the Google Groups 
"Racket Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to racket-users+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Re: [racket-users] "table" data structure in Racket

[Alex Harsanyi]

On Thursday, February 21, 2019 at 7:19:39 AM UTC+8, travis.h...@gmail.com 
wrote:
>
> Hi All,
>
> I'm resurrecting this thread to ask if anyone in the Racket community has 
> Apache Arrow on their radar. It seems like Apache Arrow might be gaining 
> steam.
>

> Apache Arrow is a cross-language development platform for in-memory data. 
>> It specifies a standardized language-independent columnar memory format for 
>> flat and hierarchical data, organized for efficient analytic operations on 
>> modern hardware. It also provides computational libraries and zero-copy 
>> streaming messaging and interprocess communication. Languages currently 
>> supported include C, C++, C#, Go, Java, JavaScript, MATLAB, Python, R, 
>> Ruby, and Rust. [Source 
>> 
>> ]
>
>
> I have no clue what it would take to make Racket a supported language. I 
> also don't have a sense of what kind of demand the Racket community has for 
> this sort of thing. Just curious if anyone is thinking about a Racket and 
> Apache Arrow pairing.
>
>
I had looked at Apache Arrow when it was first mentioned in this thread, 
but the project is large and providing bindings seemed like a large task -- 
the Java, Rust, JavaScript and Python bindings are large projects by 
themselves.  I suspect Apache Arrow (along with numpy and other scientific 
/ data-processing packages) are backed by several large companies who 
provide the development resources.

Providing usable Racket bindings for Apache Arrow is not something one can 
do as a hobby in a few weekends.  I ended up writing my own data frame 
package, which is smaller, simpler and has the features and performance 
levels that I need.

Alex. 
 

> Thanks,
>
> Travis 
>
> On Wednesday, April 6, 2016 at 1:27:56 AM UTC-7, Konrad Hinsen wrote:
>>
>> On 05/04/2016 21:12, Asumu Takikawa wrote: 
>>
>> > I haven't built anything like that, but I was hoping that we could get 
>> a GSoC 
>> > student for it (that didn't pan out though obviously). The idea was to 
>> use 
>> > packages from Python/Julia/R as inspiration: 
>> > 
>> >http://pandas.pydata.org/pandas-docs/stable/index.html 
>> >http://dataframesjl.readthedocs.org/en/latest/ 
>> >https://github.com/Rdatatable/data.table/wiki 
>>
>> Also consider this one: 
>>
>> https://arrow.apache.org/ 
>>
>> Konrad. 
>>
>

-- 
You received this message because you are subscribed to the Google Groups 
"Racket Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to racket-users+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Re: [racket-users] "table" data structure in Racket

[travis . hinkelman]

Hi All,

I'm resurrecting this thread to ask if anyone in the Racket community has 
Apache Arrow on their radar. It seems like Apache Arrow might be gaining 
steam.

Apache Arrow is a cross-language development platform for in-memory data. 
> It specifies a standardized language-independent columnar memory format for 
> flat and hierarchical data, organized for efficient analytic operations on 
> modern hardware. It also provides computational libraries and zero-copy 
> streaming messaging and interprocess communication. Languages currently 
> supported include C, C++, C#, Go, Java, JavaScript, MATLAB, Python, R, 
> Ruby, and Rust. [Source 
> 
> ]


I have no clue what it would take to make Racket a supported language. I 
also don't have a sense of what kind of demand the Racket community has for 
this sort of thing. Just curious if anyone is thinking about a Racket and 
Apache Arrow pairing.

Thanks,

Travis 

On Wednesday, April 6, 2016 at 1:27:56 AM UTC-7, Konrad Hinsen wrote:
>
> On 05/04/2016 21:12, Asumu Takikawa wrote: 
>
> > I haven't built anything like that, but I was hoping that we could get a 
> GSoC 
> > student for it (that didn't pan out though obviously). The idea was to 
> use 
> > packages from Python/Julia/R as inspiration: 
> > 
> >http://pandas.pydata.org/pandas-docs/stable/index.html 
> >http://dataframesjl.readthedocs.org/en/latest/ 
> >https://github.com/Rdatatable/data.table/wiki 
>
> Also consider this one: 
>
> https://arrow.apache.org/ 
>
> Konrad. 
>

-- 
You received this message because you are subscribed to the Google Groups 
"Racket Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to racket-users+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Re: [racket-users] "table" data structure in Racket

[Konrad Hinsen]

On 05/04/2016 21:12, Asumu Takikawa wrote:


I haven't built anything like that, but I was hoping that we could get a GSoC
student for it (that didn't pan out though obviously). The idea was to use
packages from Python/Julia/R as inspiration:

   http://pandas.pydata.org/pandas-docs/stable/index.html
   http://dataframesjl.readthedocs.org/en/latest/
   https://github.com/Rdatatable/data.table/wiki


Also consider this one:

   https://arrow.apache.org/

Konrad.

--
You received this message because you are subscribed to the Google Groups "Racket 
Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to racket-users+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Re: [racket-users] "table" data structure in Racket

['John Clements' via Racket Users]

> On Apr 5, 2016, at 12:19 PM, Jens Axel Søgaard  wrote:
> 
> Perhaps Ray Racine's library?
> 
> https://github.com/RayRacine/munger/tree/master
> 
> Look at the munger/frame/frame.rkt

I took a look, but… no docs.

Also, I like the idea of a fairly thin skin over an existing db implementation.

John


-- 
You received this message because you are subscribed to the Google Groups 
"Racket Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to racket-users+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.


signature.asc
Description: PGP signature

[racket-users] "table" data structure in Racket

['John Clements' via Racket Users]

I’ve been dabbling in data analysis in Racket for the last few months, and one 
thing that’s become immediately obvious is that we don’t have a nice notion of 
a “table”.

I believe that the natural notion of a “table” here is more or less the same in 
R and in databases: A table consists of a set of column names, and a set of 
tuples with values for each of those columns.

After building my own and messing with it for a while, I basically rewrote it 
as a front-end to sqlite3. So, for instance, you can create a table by writing

(make-table
  ‘(student qtr score)
  ‘((“janice” 2152 32.4) (“franklin” 2164 29.9) (“zongthar” 2174 .9))

… and then query it in a variety of useful ways.

1) has anyone else already built and released this, and I just couldn’t find it 
in pkgs?

2) Is it worth building the functional-skinned interface, or is it easier just 
to write lots of SQL code?

John



-- 
You received this message because you are subscribed to the Google Groups 
"Racket Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to racket-users+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Re: [racket-users] "table" data structure in Racket

[Jens Axel Søgaard]

Perhaps Ray Racine's library?

https://github.com/RayRacine/munger/tree/master

Look at the munger/frame/frame.rkt

2016-04-05 21:04 GMT+02:00 'John Clements' via Racket Users <
racket-users@googlegroups.com>:

> I’ve been dabbling in data analysis in Racket for the last few months, and
> one thing that’s become immediately obvious is that we don’t have a nice
> notion of a “table”.
>
> I believe that the natural notion of a “table” here is more or less the
> same in R and in databases: A table consists of a set of column names, and
> a set of tuples with values for each of those columns.
>
> After building my own and messing with it for a while, I basically rewrote
> it as a front-end to sqlite3. So, for instance, you can create a table by
> writing
>
> (make-table
>   ‘(student qtr score)
>   ‘((“janice” 2152 32.4) (“franklin” 2164 29.9) (“zongthar” 2174 .9))
>
> … and then query it in a variety of useful ways.
>
> 1) has anyone else already built and released this, and I just couldn’t
> find it in pkgs?
>
> 2) Is it worth building the functional-skinned interface, or is it easier
> just to write lots of SQL code?
>
> John
>
>
>
> --
> You received this message because you are subscribed to the Google Groups
> "Racket Users" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to racket-users+unsubscr...@googlegroups.com.
> For more options, visit https://groups.google.com/d/optout.
>



-- 
-- 
Jens Axel Søgaard

-- 
You received this message because you are subscribed to the Google Groups 
"Racket Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to racket-users+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.