Saturday, March 18, 2017

Simple CSV Transformations

Here's an interesting question:

I came across your blog post "Introduction to using Python to process CSV files" as I'm looking to do something I'd think is easy in Python but I don't know how to do it. 

I simply want to examine a column then create a new column based on an if-then on the original column. So if my CSV has a "gender" field I'd like to do the Python equivalent of this SQL statement: 

case when gender = 'M' then 1 else 0 end as gender_m, case when gender = 'F' then 1 else 0 end as gender_f,...

I can do it in Pandas but my CSVs are too big and I run into memory issues. 

There are a number of ways to tackle this.

First -- and foremost -- this is almost always just one step in a much longer and more complex set of operations. It's a little misleading to read-and-write a CSV file to do this.

A little misleading.

It's not wrong to write a file with expanded data. But the "incrementally write new files" process can become rather complex. If we have a large number of transformations, we can wind up with many individual file-expansion steps. These things often grow organically and can get out of control. A complex set of steps should probably be collapsed into a single program that handles all of the expansions at once.

This kind of file-expansion is simple and fast. It can open a door previously closed by the in-memory problem  of trying to do the entire thing in pandas.

The general outline looks like this

from pathlib import Path
import csv
source_path = Path("some_file.csv")
target_path = Path(source_path.stem + "_1").with_suffix('.csv')

def transform(row):
    return row

with as source_file:
    with'w', newline='') as target_file:
        reader = csv.DictReader(source_file)
        columns =  reader.fieldnames + ['gender_m', 'gender_f']
        writer = csv.DictWriter(target_file, columns)
        for row in reader:
            new_row = transform(row)

The goal is to be able put some meaningful transformation processing in place of the build new_row comment.

The overall approach is this.

1. Create Path objects to refer to the relevant files.

2. Use with-statement context managers to handle the open files. This assures that the files are always properly closed no matter what kinds of exceptions are raised.

3. Create a dictionary-based reader for the input.  Add the additional columns and create a dictionary-based writer for the output. This allows the processing to work with each row of data as a dictionary.
This presumes that the data file actually has a single row of heading information with column names.

If column names are missing, then a fieldnames attribute can be provided when creating the DictReader(), like this: csv.DictReader(source_file, ['field', 'field', ...]).

The for statement works because a csv Reader is an iterator over each row of data.

I've omitted any definition of the transformational function. Right now, it just returns each row unmodified. We'd really like it to do some useful work.

Building The New Row

The transformation function needs to build a new row from an existing row.

Each row will be a Python dictionary. A dictionary is a mutable object. We aren't really building a completely new object -- that's a waste of memory. We'll modify the row object, and return it anyway. It will involve a microscopic redundancy of creating two references to the same dictionary object, one known by the variable name row and the other know by new_row.

Here's an example body for transform()

def transform(row):
    row['gender_m'] = 1 if row['gender'] == 'M' else 0
    row['gender_f'] = 1 if row['gender'] == 'F' else 0
    return row

This will build two new keys in the row dictionary. The exact two keys added to the fieldnames to write a new file.

Each key be associated with a value computed by a simple expression. In this case, the logical if-else operator is used to map a boolean value, row['gender'] == 'M', to one of two integer values, 1 or 0.

If this is confusing -- and it can be -- this can also be done with if statements instead of expressions.

def transform(row):
    if row['gender'] == 'M':
        row['gender_m'] = 1
        row['gender_m'] = 0
    row['gender_f'] = 1 if row['gender'] == 'F' else 0
    return row

I only rewrite the 'M' case. I'll leave the rewrite of the 'F' case to the reader.

Faster Processing with a Generator

We can simplify the body of the script slightly. This will make it work a hair faster. The following statements involve a little bit of needless overhead.

        for row in reader:
            new_row = transform(row)

We can change this as follows:

        writer.writerows(transform(row) for row in reader)

This uses a generator expression, transform(row) for row in reader, to build individually transformed rows from a source of data. This doesn't involve executing two statements for each row of data. Therefore, it's faster.

We can also reframe it like this.

        writer.writerows(map(transform, reader))

In this example, we've replaced the generator expression with the map() function. This applies the transform() function to each row available in the reader.

In both cases, the writer.writerows() consumes the data produced by the generator expression or the map() function to create the output file.

The idea is that we can make the transform() function as complex as we need. We just have to be sure that all the new field names are handled properly when creating the writer object.

Tuesday, March 14, 2017

Strange "I Hate Python" Concerns. Mostly nonsensical.

From a troll-bait thread asking the argumentative question "why do people use Python"? The answers were, oddly, a long list of strange, nonsensical complaints. And a few logical fallacies. Here are all the ones I could parse:
  1. "It's the FORTRAN of our times."
  2. It's the COBOL of our times.
  3. "deep seated aversion to languages where whitespace has fundamental syntactic significance". 
  4. "And also where the simplest "Hello world!" program is busted between v2 and v3 (true story)"
  5. "My stomach turns in a knot at the introduction of EVERY trendy language"
  6. "I am almost always focused on productization qualities such as maintainability, performance, and any number of other "-ilities"."
  7. Nobody [cares] about Your language unless You can produce executable
  8. "It's ghastly. The Python Tools for Visual Studio eases the pain with a full symbolic/visual debugger but still..."
  9. "the socialist theme of universities leads to preference for open source and "free" over professionally developed and maintained tools... Meanwhile I really like JavaScript as a free wheeling scripting language."
  10. "Python ... is an inferior language. I can trust a well-engineered JavaScript system."
  11. "it's worse than fortran because it has a dedicated following"
  12. "my indictment is maintainability once productized. I always have a fear of building legacy packages that, once a mountain is built and is difficult to move, that people of the future will curse my name"
  13. "rationally, the continuing investment in the Node/TypeScript infrastructure places JavaScript in an entirely different infrastructure realm than Python"
  14. "Python doesn't have its equivalent of Node.js"
  15. "as a LANGUAGE JavaScript has great infrastructure across device types, OS brands, and across every level of scale now imaginable"
  16. Four separate reasoning-by-analogy: Lisp, FoxPro, PHP, and Perl. (e.g., "Amazon did amazing things with perl.") Somehow a failure involving these languages (or ecosystems or whatever) indicts Python because they're all "trendy" (I think.)
Yes. There were others that made less sense. I've omitted them.

TL;DR: These people don't seem to know what they're talking about. The posts are universally fact-free, so we're just listening to folks rambling randomly about Python.

Some responses. Feel free to use these when someone asks you why you're still using Python.
  1. That makes no sense
  2. That makes no sense
  3. The languages which are totally free of whitespace appear to be C and maybe C++. This principle rules out JavaScript, since the ASI rules involve wrangling ";" in place of the evil whitespace.
  4. This is a weird complaint. Stuff changed. How is that a problem? Are you saying change is a problem? What's this then? 
  5. Trendy is a problem? Really?
  6. Who isn't focused on quality attributes?
  8. What does "ghastly" mean?
  9. What's a "socialist theme"? How is JavaScript "free-wheeling"? What does that even mean?
  10. What is "inferior" being measured? Alphabetically? (Python comes after Javascript, so it's in an inferior position alphabetically?)
  11. How is a dedicated following a problem?
  12. Python is second to Java.
  13. "continuing investment"? By whom? And how does this "investment" compare with Python?
  14. What's wrong with twisted, tornado, Gunicorn, and Nginx? Don't they count?
  15. Python is available more-or-less everywhere. Without a specific coverage gap, this makes no sense.
  16. Also known as the False Equivalence fallacy. Without details of the failure mode, equivalence with Python isn't established.
Omitted is a random discussion on how Ruby is "rigorously defined". The implication seems to be that Python somehow might not be rigorously defined or something. It's not clear what the sub-thread was about, so I ignored it.

This thread seemed to involve two kinds of complaints:
  • Utter nonsense.
  • Lies that are pretty east to refute.

Tuesday, March 7, 2017

Chain of Command Example

One objective of the Chain of Command design pattern is to be able to write a bunch of functions that link together. The form a chain of alternative implementations. The idea is the have alternatives that vary in their ability to compute a correct answer. If algorithm 1 doesn’t work, try algorithm 2. If that doesn’t work, fall back to algorithm 3, etc.

Perhaps algorithm 1 has a number of constrains: it's fast, but only for a limited kind of input. Algorithm 2 may have a different set of constraints. And Algorithm 3 involves the "British Museum" algorithm.

Algorithm zero, at the head of the chain, can be a dynamic cache, perhaps with LRU features. Maybe it can be shared among servers. There are lots of choices here. The idea is that a cache is often first because it's so fast.

We can, of course, write a giant master function with other functions. Maybe they're all linked with a lot of clever if-statements. We know how that turns out, don't we?

Instead, we can make each function a distinct object. The alternative algorithm functions have a relationship with other functions, so a simple non-stateful class definition is appropriate. The cache alternative may involve state changes, so it’s a little different than the others.

We'll imagine a simple doTheThing() function with a few arguments that returns a value. We have several alternatives. The goal to be able to wrap each doTheThing() function in a very small class like this:

class AlgorithmOne(DoAThing):
    """One way to do it."""
    def doTheThing(self, arg1, arg2):
        # Check some constraints, maybe...
        if arg1 < arg2:
            return Fraction(arg1, arg2)
            raise DoAThingError("Outside AlgorithmOne Constraints")

Either this algorithm produces a good answer, or it raises an exception that it can’t really do the thing. Any other exceptions are ordinary bad code crashing at run time.

The core feature of the abstract superclass is the all-important try:/except: block that tries doTheThing(). If the DoAThingError exception is raised, it moves down the chain of command. If it succeeds, then, we're done.

This has a consequence of wrapping the doTheThing() implementation with a function named theThing(). The wrapper function, theThing(), contains the try:/except: block, a call to a concrete doTheThing() implementation, plus the fall-back processing.

The cache version doesn't really have a meaningful implementation of the theThing() function. Instead it always tries the fallback chain and caches the result.

A cool way to build the chain is omitted from this design. We're creating a linked list like this:

a2 = AlgorithmTwo()
a1 = AlgorithmOne(a2)
coc = UseCache(a1)

Some people object to the "backwardness" of this. In which case, they can write a simple constructor function which emits the chain of command by linking the things together in the proper order.

def builder(*classes):
    previous = None
    for class_ in classes:
        next = class_(previous)
        previous = next
    return previous

I'm not sure it's essential. But it's simple.

Here's the whole show.

"""Chain of command."""
from fractions import Fraction
import pickle

class DoAThingError(Exception):

class DoAThing:
    """Abstract superclass."""
    def __init__(self, fall_back=None):
        self.fall_back = fall_back

    def theThing(self, arg1, arg2):
            return self.doTheThing(arg1, arg2)
        except DoAThingError:
            if self.fall_back:
                return self.fall_back.theThing(arg1, arg2)

    def doTheThing(self, arg1, arg2):
        raise DoAThingError("Not Implemented")

class UseCache(DoAThing):
    """Is the answer in cache? Cache is dynamic and grows quickly.
    There's no LRU.
    def __init__(self, *args, **kw):
        super().__init__(*args, **kw)
        self.cache = {}

    def load(self, openFile):
        self.cache = pickle.load(openFile)

    def dump(self, openFile):
        pickle.dump(self.cache, openFile)

    def theThing(self, arg1, arg2):
        if (arg1, arg2) not in self.cache:
            self.cache[arg1, arg2] = self.fall_back.theThing(arg1, arg2)
        return self.cache[arg1, arg2]

class AlgorithmOne(DoAThing):
    """One way to do it."""
    def doTheThing(self, arg1, arg2):
        # Check some constraints, maybe...
        if arg1 < arg2:
            return Fraction(arg1, arg2)
            raise DoAThingError("Outside AlgorithmOne Constraints")

class AlgorithmTwo(DoAThing):
    """Another way to do it."""
    def doTheThing(self, arg1, arg2):
        return arg1/arg2

a2 = AlgorithmTwo()
a1 = AlgorithmOne(a2)
coc = UseCache(a1)


Tuesday, February 14, 2017

Intro to Python CSV Processing for Actual Beginners

I've written a lot about CSV processing. Here are some examples

It crops up in my books. A lot.

In all cases, though, I make the implicit assumption that my readers already know a lot of Python. This is a disservice to anyone who's getting started.

Getting Started

You'll need Python 3.6. Nothing else will do if you're starting out.

Go to and get Python 3.6. You can get the small "miniconda" version to start with. It has some of what you'll need to hack around with CSV files. The full Anaconda version contains a mountain of cool stuff, but it's a big download.

Once you have Python installed, what next? To be sure things are running do this:
  1. Find a command line prompt (terminal window, cmd.exe, whatever it's called on your OS.)
  2. Enter python3.6 (or just python in Windows.)
  3. If Anaconda installed everything properly, you'll have an interaction that looks like this:

MacBookPro-SLott:Python2v3 slott$ python3.5
Python 3.5.1 (v3.5.1:37a07cee5969, Dec  5 2015, 21:12:44) 
[GCC 4.2.1 (Apple Inc. build 5666) (dot 3)] on darwin
Type "help", "copyright", "credits" or "license" for more information.

More-or-less. (Yes, the example shows 3.5.1 even though I said you should get 3.6. As soon as the course drops, I'll upgrade. The differences between 3.5 and 3.6 are almost invisible.)

Here's your first interaction.

>>> 355/113

Yep. Python did math. Stuff is happening.

Here's some more.

>>> exit
Use exit() or Ctrl-D (i.e. EOF) to exit
>>> exit()

Okay. That was fun. But it's not data wrangling. When do we get to the good stuff?

To Script or Not To Script

We have two paths when it comes to scripting. You can write script files and run them. This is pretty normal application development stuff. It works well. 


You can use a Jupyter Notebook. This isn't exactly a script. But. You can use it like a script. It's a good place to start building some code that's useful. You can rerun some (or all) of the notebook to make it script-like.

If you downloaded Anaconda, you have Jupyter. Done. Skip over the next part on installing Jupyter.

Installing Jupyter

If you did not download the full Anaconda -- perhaps because you used the miniconda -- you'll need to add Jupyter.  You can use the command conda install jupyter for this.

Another choice is to use the PIP program to install jupyter. The net effect is the same. It starts like this

MacBookPro-SLott:Python2v3 slott$ pip3 install jupyter
Collecting jupyter
  Downloading jupyter-1.0.0-py2.py3-none-any.whl
Collecting ipykernel (from jupyter)
  Downloading ipykernel-4.5.2-py2.py3-none-any.whl (98kB)

    100% |████████████████████████████████| 102kB 1.3MB/s 

It ends like this.

  Downloading pyparsing-2.1.10-py2.py3-none-any.whl (56kB)
    100% |████████████████████████████████| 61kB 2.1MB/s 
Installing collected packages: ipython-genutils, decorator, traitlets, appnope, appdirs, pyparsing, packaging, setuptools, ptyprocess, pexpect, simplegeneric, wcwidth, prompt-toolkit, pickleshare, ipython, jupyter-core, pyzmq, jupyter-client, tornado, ipykernel, qtconsole, terminado, nbformat, entrypoints, mistune, pandocfilters, testpath, bleach, nbconvert, notebook, widgetsnbextension, ipywidgets, jupyter-console, jupyter
  Found existing installation: setuptools 18.2
    Uninstalling setuptools-18.2:
      Successfully uninstalled setuptools-18.2
  Running install for simplegeneric ... done
  Running install for tornado ... done
  Running install for terminado ... done
  Running install for pandocfilters ... done
Successfully installed appdirs-1.4.0 appnope-0.1.0 bleach-1.5.0 decorator-4.0.11 entrypoints-0.2.2 ipykernel-4.5.2 ipython-5.2.2 ipython-genutils-0.1.0 ipywidgets-5.2.2 jupyter-1.0.0 jupyter-client-4.4.0 jupyter-console-5.1.0 jupyter-core-4.2.1 mistune-0.7.3 nbconvert-5.1.1 nbformat-4.2.0 notebook-4.4.1 packaging-16.8 pandocfilters-1.4.1 pexpect-4.2.1 pickleshare-0.7.4 prompt-toolkit-1.0.13 ptyprocess-0.5.1 pyparsing-2.1.10 pyzmq-16.0.2 qtconsole-4.2.1 setuptools-34.1.1 simplegeneric-0.8.1 terminado-0.6 testpath-0.3 tornado-4.4.2 traitlets-4.3.1 wcwidth-0.1.7 widgetsnbextension-1.2.6

Now you have Jupyter.

What just happened? You installed a large number of Python packages. All of those packages were required to run Jupyter. You can see jupyter-1.0.0 hidden in the list of packages that were installed.

Starting Jupyter

The Jupyter tool does a number of things. We're going to use the notebook feature to save some code that we can rerun. We can also save notes and do other things in the notebook. When you start the notebook, two things will happen.
  1. The terminal window will start displaying the Jupyter console log.
  2. A browser will pop open showing the local Jupyter notebook home page.
Here's what the console log looks like:

MacBookPro-SLott:Python2v3 slott$ jupyter notebook
[I 08:51:56.746 NotebookApp] Writing notebook server cookie secret to /Users/slott/Library/Jupyter/runtime/notebook_cookie_secret
[I 08:51:56.778 NotebookApp] Serving notebooks from local directory: /Users/slott/Documents/Writing/Python/Python2v3
[I 08:51:56.778 NotebookApp] 0 active kernels 
[I 08:51:56.778 NotebookApp] The Jupyter Notebook is running at: http://localhost:8888/?token=2eb40fbb96d7788dd05a49600b1fca4e07cd9c8fe931f9af
[I 08:51:56.778 NotebookApp] Use Control-C to stop this server and shut down all kernels (twice to skip confirmation).

You can glance at it to see that things are still working. The "Use Control-C to stop this server" is a reminder of how to stop things when you're done.

Your Jupyter home page will have this logo in the corner. Things are working.

You can pick files from this list and edit them. And -- important for what we're going to do -- you can create new notebooks.

On the right side of the web page, you'll see this:

You can create files and folders. That's cool. You can create an interactive terminal session. That's also cool. More important, though, is that you can create a new Python 3 notebook. That's were we'll wrangle with CSV files.

"But Wait," you say. "What directory is it using for this?"

The jupyter server is using the current working directory when you started it.

If you don't like this choice, you have two alternatives.
  • Stop Jupyter. Change directory to your preferred place to keep files. Restart Jupyter.
  • Stop Jupyter. Include the --notebook-dir=your_working_directory option.
The second choice looks like this:

MacBookPro-SLott:Python2v3 slott$ jupyter notebook --notebook-dir=~/Documents/Writing/Python
[I 11:15:42.964 NotebookApp] Serving notebooks from local directory: /Users/slott/Documents/Writing/Python

Now you know where your files are going to be. You can make sure that your .CSV files are here. You will have your ".ipynb" files here also. Lots of goodness in the right place.

Using Jupyter

Here's what a notebook looks like. Here's a screen shot.

First. The notebook was originally called "untitled" which seemed less than ideal. So I clicked on the name and changed it to "csv_wrestling".

Second. There was a box labeled In [ ]:. I entered some Python code to the right of this label. Then I clicked the run cell icon. (It's similar to this emoji --  ⏯ -- but not exactly.)

The In [ ]: changed to In [1]:. A second box appeared labeled Out [1]:. This annotates our dialog with Python: each input and Python's response is tracked. It's pretty nice. We can change our input and rerun the cell. We can add new cells with different things to run. We can run all of the cells. Lots of things are possible based on this idea of a cell with our command. When we run a cell, Python processes the command and we see the output.

For many expressions, a value is displayed.  For some expressions, however, nothing is displayed. For complete statements, nothing is displayed. This means we'll often have to throw the name of a variable in to see the value of that variable.

The rest of the notebook is published separately. It's awkward to work in Blogger when describing a Jupyter notebook. It's much easier to simply post the notebook in GitHub.

The notebook is published here: slott56/introduction-python-csv. You can follow the notebook to build your own copy which reads and writes CSV files.

Tuesday, February 7, 2017

Writing Tools

Read this:

What I have been doing instead of using these sophisticated, integrated writing tools?

I use OmniOutliner. I've used it for years. It does a lot of things. Most notably, I can create multiple columns so that I can create page budgets for outlines. Acquisition Editors like this. Except, of course, they like it as an DOCX file, which requires a bit of manual juggling to produce.

I use BBEdit and KomodoEdit for a the bulk of my writing.

"But wait," you say, "those are text editors."

(Or, more dismissively, "there are merely text editors.")

Correct.  I use RST markup and write in Unicode text.  I use tools to convert the RST text to a variety of other binary formats. See for a list of tools. This is what I often use:

How is this better than a tool like Scrivener? It depends -- as always -- on what you're trying to optimize. My pipeline has the (dubious) advantage of being very inexpensive. Except for OmniOutliner and BBEdit, it's all community-edition, free software. If cheap is your goal, I've got cheap.

The cool part is this.

The Mac OS X desktop is an integrated writing environment. I have browser, outliner, writing tool, publishing tool, etc., etc., all readily and immediately available. The "look and feel" isn't consistent, but I'm not sure that's a show-stopper.

The biggest difficulty?

BBEdit doesn't enable the Mac OS X grammar checker. Really. It's switched off. The grammar checker is sometimes handy for preventing a large number of common, dumb writing mistakes. BBEdit shows the word count, which is very helpful for some kinds of writing. I wind up using a second app (i.e. the built-in Mac OS X TextEdit) to make a grammar check pass.

I think, however, the hacker-friendly free-and-open-source tool chain may have reached the end of its service life.

Why Not Use Word?

"After all," you say, "MS-Word does everything."

Agreed. It does everything badly and confusingly. (1) The outliner is hard to use and is firmly tied to the text in a way that breaks outlines all the time. (What's that paragraph doing there? Why is it the wrong outline level?) (2) There are too many useless features. The presence of "advanced" mode is a UX nightmare come true. (3) The character-mode and paragraph-mode formatting rules are baffling (and break the outlining.) (4) The styles are essentially invisible: you have to click on the text and check the style side-bar to be sure that the (invisible) markup is actually right.

The worst thing is that publishers have house style sheets for MS-Word that drive the publishing pipeline. This means that writing involves a weird step where I have to apply the publishers styles to things that are **very** clearly annotated with RST markup. You have to review each word. The words may look right, but have the wrong style applied. This is extremely tiresome to get right.

I intend to stick with plain-text markup. Scrivener supports MultiMarkdown. It's not RST, but it seems to be as rich with built-in semantic categories.