photo

Rick Amos

MathWorks

Last seen: 3 months ago |  Active since 2012

Followers: 0   Following: 0

Message

Developer at MathWorks, working on Parallel Computing Toolbox.

Statistics

All
  • Knowledgeable Level 2
  • First Answer
  • Solver

View badges

Feeds

View by

Answered
Reading portions of several CSV files
You might want to take a look at tabularTextDatastore and/or tall arrays. These are geared up to do exactly this kind of thing:-...

5 years ago | 5

Answered
tallArray sorting and plot error
Instead of converting State to a character vector, I would recommend using one of the string utilities. These are all supported ...

5 years ago | 0

Answered
Pass fileDatastore explicit AWS credentials
It is currently not possible to pass explicit credentials directly into fileDatastore. Datastore retrieves the credentials in th...

5 years ago | 0

Answered
Too many .mat files when using Write Tall Table
As you've likely guessed, vertical concatenation of 1600 arrays results in at least 1600 files from tall/write. This is the case...

5 years ago | 0

| accepted

Answered
Generate a Datastore for binary data?
As of R2017b, MATLAB now allows you to write custom datastores. This will allow you to create datastores that handle large binar...

7 years ago | 2

Answered
How to check Hadoop and Matlab Integrated Properly or Not ???
To look at whether MATLAB is running on the Hadoop cluster correctly, your best bet is to look at the Hadoop/Yarn Web UI. By def...

7 years ago | 0

| accepted

Answered
How to create tall datastore from multiple data parts?
A datastore can be created from a collection of folders and so the easiest way to achieve this is to place each block of data in...

8 years ago | 0

Answered
Datastore for custom files
For files on Hadoop, the File Datastore makes a temporary copy of the file on the local drive to be read by the ReadFcn. The Rea...

8 years ago | 0

| accepted

Answered
How to solve ' All the input files must be Sequence files. Invalid file: '' '
This error message occurred because MATLAB could not find the output files generated by the Hadoop Job. For now, this error mess...

9 years ago | 1

| accepted

Answered
How to solve" The user home directory '/homes/' for the cluster either did not exist or was not writable by the HADOOP user."
The '/homes/' folder is related to the Hadoop property 'mapreduce.admin.user.home.dir'. This is used by Hadoop to set the HOME e...

9 years ago | 1

| accepted

Answered
How to make a MAT-file that can be used to create a Datastore for MapReduce?
In R2014b there is currently not a direct way of creating a MAT file datastore. However, there are several indirect ways that wi...

10 years ago | 2

| accepted

Answered
Mapreduce does not seem to use all available cores
In R2014b, there are some limitations with the minimum size of data that can be parallelized. To avoid this limitation, the inpu...

10 years ago | 0

| accepted

Answered
MCR problems with linux/hadoop
This behavior arises because Hadoop defines the 'HOME' environment variable of task processes to be '/homes' by default. MATLAB ...

10 years ago | 0