Using Pandas to create, read, and update hdf5 file structure - python

We would like to be able to allow the HDF5 files themselves to define their columns, indexes, and column types instead of maintaining a separate file that defines structure of the HDF5 data.
How can I create an empty HDF5 file from Pandas with a specific table structure like:
Columns
id (Int)
name (Str)
update_date (datetime)
some_float (float)
Indexes
id
name
Once the HDF5 is created and saved to disk, how do I retrieve the column and index information without having to open the file completely each time since it will likely contain several GB of data.
Many thanks in advance...
-- UPDATE --
Thanks for the comments. To clarify a bit more:
We do have some experience with Pandas but by no means are really proficient. The part that is tripping us up is creating an empty data structure and reading that structure from a file that you will not want to fully open. In all of the Pandas examples there is data. The Pandas examples also only show two ways to retrieve data/structure which are to read the entire frame into memory or issue a where clause. In this case, we would like to be able to see the table structure without query operations if possible.
I know this is an odd case. Why the heck would you want an empty dataframe?? Well, we want to have a great deal of flexility in moving data around and want to be able to define a target dataframe structure prior to data writing, which could take place much later (e.g. hours or days). Since the HDF5 specification maintains all that information it seems directionally incorrect to store the table structure information separately. Thus our desire to crack the code on this subject.
-- UPDATE 2 --
To add more detail as #jeff requested.
We would like to abstract some of the common Pandas functions like summing data or merging two frames. Thus we would like to be able to ask each frame what their columns are so we can present a view for the user to select the result frame columns.
For example, if we imported a CSV with columns A, B, C, D, and V and saved the frame to HDF5 as my_csv.hdf then we would be able to determine the columns by opening the file.
However, in our use case it is likely that the import frame for the CSV could be cleared periodically and no longer contain the data. The reason knowing that the my_csv frame has certain columns and types is important because we want to enable a user to then select those columns for summing in a downstream operation. Lets say a user wants to sum column V by the values in columns A and B only and save the frame as my_sum. Since we can't ensure my_csv will always have data we would like to ensure it at least contains the structure.
Open to other suggestions obviously. It is also possible to store the table structure info in the user_block. This, again, is not ideal because the structure is now being kept in two different areas but I guess it would be possible to always update the user_block on save using the latest column and index information for the frame, although I believe the to_* operations in Pandas will blow away the user_block so...blah. I feel like I'm talking myself into maintaining a peer structure definition but I REALLY would love some suggestions to not have to do that.

Related

How can I pass through dataset metadata, like a hash or timestamp, in a Python Transform in Palantir Foundry?

I want to be able to do two things:
Store a hash of a datasets contents (so I can decide whether it has updated). To date, I have done this via a second output dataset with a single row that stores the hash and row count. In my Transform I can read that output and compare it to the current build's hash and row count to decide if data has updated. This works fine, but I'd like to avoid having a second dataset if possible.
Pass through timestamps from upstream dependencies so that in downstream workflows I can answer "when did dependency X last update?"
It seems like both of these could be solved by some sort of key-value metadata store on the dataset.
You're correct that one of the most straightforward ways to do this is to decorate the rows with a timestamp value, and in fact with Foundry's Parquet storage system, this will be encoded using Dictionary Encoding, a highly efficient mechanism to store repeated values.
The problem with this approach is you'll have to stack a new column for each phase of updating you want to keep track of. This might prove annoying to maintain in practice.
However, if you don't want to add this data to your rows and instead simply want to store your metadata, you have two options, one of which you've already found:
Store metadata in a separate dataset
Write an 'unused' file (probably .csv or .txt) to your output keeping track of this information
Foundry won't consider your .csv or .txt extra file on the output if you're writing a standard DataFrame to it since your schema by default will only read Parquet files. This means you can store this little snippet of information without affecting your output. If you check platform documentation, you can confirm that it's possible to write both a DataFrame to an output and a file of your own.
It may be simpler to interact with a second output however since the mechanisms of Incremental Transforms and schema handling will be taken care of for you, so I'd recommend proceeding with 1. as you are right now.

Pandas - retrieving HDF5 columns and memory usage

I have a simple question, I cannot help but feel like I am missing something obvious.
I have read data from a source table (SQL Server) and have created an HDF5 file to store the data via the following:
output.to_hdf('h5name', 'df', format='table', data_columns=True, append=True, complib='blosc', min_itemsize = 10)
The dataset is ~50 million rows and 11 columns.
If I read the entire HDF5 back into a dataframe (through HDFStore.select or read_hdf), it consumes about ~24GB of RAM. If I parse specific columns into the read statements (e.g. selecting 2 or 3 columns), the dataframe now only returns those columns, however the same amount of memory is consumed (24GB).
This is running on Python 2.7 with Pandas 0.14.
Am I missing something obvious?
EDIT: I think I answered my own question. While I did a ton of searching before posting, obviously once posted I found a useful link: https://github.com/pydata/pandas/issues/6379
Any suggestions on how to optimize this process would be great, due to memory limitations I cannot hit peak memory required to release via gc.
HDFStore in table format is a row oriented store. When selecting the query indexes on the rows, but for each row you get every column. selecting a subset of columns does a reindex at the end.
There are several ways to approach this:
use a column store, like bcolz; this is currently not implemented by PyTables so this would involve quite a bit of work
chunk thru the table, see here and concat at the end - this will use constant memory
store as a fixed format - this is a more efficient storage format so will use less memory (but cannot be appended)
create your own column store-like by storing to multiple sub tables and use select_as_multiple see here
which options you choose depend on the nature of your data access
note: you may not want to have all of the columns as data_columns unless you are really going to select from the all (you can only query ON a data_column or an index)
this will make store/query faster

HDF5 Links to Events in Dataset

I'm trying to use HDF5 to store time-series EEG data. These files can be quite large and consist of many channels, and I like the features of the HDF5 file format (lazy I/O, dynamic compression, mpi, etc).
One common thing to do with EEG data is to mark sections of data as 'interesting'. I'm struggling with a good way to store these marks in the file. I see soft/hard links supported for linking the same dataset to other groups, etc -- but I do not see any way to link to sections of the dataset.
For example, let's assume I have a dataset called EEG containing sleep data. Let's say I run an algorithm that takes a while to process the data and generates indices corresponding to periods of REM sleep. What is the best way to store these index ranges in an HDF5 file?
The best I can think of right now is to create a dataset with three columns -- the first column is a string and contains a label for the event ("REM1"), and the second/third column contains the start/end index respectively. The only reason I don't like this solution is because HDF5 datasets are pretty set in size -- if I decide later that a period of REM sleep was mis-identified and I need to add/remove that event, the dataset size would need to change (and deleting the dataset/recreating it with a new size is suboptimal). Compound this by the fact that I may have MANY events (imagine marking eyeblink events), this becomes more of a problem.
I'm more curious to find out if there's functionality in the HDF5 file that I'm just not aware of, because this seems like a pretty common thing that one would want to do.
I think what you want is a Region Reference — essentially, a way to store a reference to a slice of your data. In h5py, you create them with the regionref property and numpy slicing syntax, so if you have a dataset called ds and your start and end indexes of your REM period, you can do:
rem_ref = ds.regionref[start:end]
ds.attrs['REM1'] = rem_ref
ds[ds.attrs['REM1']] # Will be a 1-d set of values
You can store regionrefs pretty naturally — they can be attributes on a dataset, objects in a group, or you can create a regionref-type dataset and store them in there.
In your case, I might create a group ("REM_periods" or something) and store the references in there. Creating a "REM_periods" dataset and storing the regionrefs there is reasonable too, but you run into the whole "datasets tend not to be variable-length very well" thing.
Storing them as attrs on the dataset might be OK, too, but it'd get awkward if you wanted to have more than one event type.

How can I ensure unique rows in a large HDF5

I'm working on implementing a relatively large (5,000,000 and growing) set of time series data in an HDF5 table. I need a way to remove duplicates on it, on a daily basis, one 'run' per day. As my data retrieval process currently stands, it's far easier to write in the duplicates during the data retrieval process than ensure no dups go in.
What is the best way to remove dups from a pytable? All of my reading is pointing me towards importing the whole table into pandas, and getting a unique- valued data frame, and writing it back to disk by recreating the table with each data run. This seems counter to the point of pytables, though, and in time I don't know that the whole data set will efficiently fit into memory. I should add that it is two columns that define a unique record.
No reproducible code, but can anyone give me pytables data management advice?
Big thanks in advance...
See this releated question: finding a duplicate in a hdf5 pytable with 500e6 rows
Why do you say that this is 'counter to the point of pytables'? It is perfectly possible to store duplicates. The user is responsible for this.
You can also try this: merging two tables with millions of rows in python, where you use a merge function that is simply drop_duplicates().

Generate table schema inspecting Excel(CSV) and import data

How would I go around creating a MYSQL table schema inspecting an Excel(or CSV) file.
Are there any ready Python libraries for the task?
Column headers would be sanitized to column names. Datatype would be estimated based on the contents of the spreadsheet column. When done, data would be loaded to the table.
I have an Excel file of ~200 columns that I want to start normalizing.
Use the xlrd module; start here. [Disclaimer: I'm the author]. xlrd classifies cells into text, number, date, boolean, error, blank, and empty. It distinguishes dates from numbers by inspecting the format associated with the cell (e.g. "dd/mm/yyyy" versus "0.00").
The job of programming some code to wade through user-entered data to decide on what DB datatype to use for each column is not something that can be easily automated. You should be able to eyeball the data and assign types like integer, money, text, date, datetime, time, etc and write code to check your guesses. Note that you need to able to cope with things like numeric or date data entered in text fields (can look OK in the GUI). You need a strategy to handle cells that don't fit the "estimated" datatype. You need to validate and clean your data. Make sure you normalize text strings (strip leading/trailing whitespace, replace multiple whitespaces by a single space. Excel text is (BMP-only) Unicode; don't bash it into ASCII or "ANSI" -- work in Unicode and encode in UTF-8 to put it in your database.
Quick and dirty workaround with phpmyadmin:
Create a table with the right amount of columns. Make sure the data fits the columns.
Import the CSV into the table.
Use the propose table structure.
As far as I know, there is no tool that can automate this process (I would love for someone to prove me wrong as I've had this exact problem before).
When I did this, I came up with two options:
(1) Manually create the columns in the db with the appropriate types and then import, or
(2) Write some kind of filter that could "figure out" what data types the columns should be.
I went with the first option mainly because I didn't think I could actually write a program to do the type inference.
If you do decide to write a type inference tool/conversion, here are a couple of issues you may have to deal with:
(1) Excel dates are actually stored as the number of days since December 31st, 1899; how does one infer then that a column is dates as opposed to some piece of numerical data (population for example)?
(2) For text fields, do you just make the columns of type varchar(n) where n is the longest entry in that column, or do you make it an unbounded char field if one of the entries is longer than some upper limit? If so, what's a good upper limit?
(3) How do you automatically convert a float to a decimal with the correct precision and without loosing any places?
Obviously, this doesn't mean that you won't be able to (I'm a pretty bad programmer). I hope you do, because it'd be a really useful tool to have.
Just for (my) reference, I documented below what I did:
XLRD is practical, however I've just saved the Excel data as CSV, so I can use LOAD DATA INFILE
I've copied the header row and started writing the import and normalization script
Script does: CREATE TABLE with all columns as TEXT, except for Primary key
query mysql: LOAD DATA LOCAL INFILE loading all CSV data into TEXT fields.
based on the output of PROCEDURE ANALYSE, I was able to ALTER TABLE to give columns the right types and lengths. PROCEDURE ANALYSE returns ENUM for any column with few distinct values, which is not what I needed, but I found that useful later for normalization. Eye-balling 200 columns was a breeze with PROCEDURE ANALYSE. Output from PhpMyAdmin propose table structure was junk.
I wrote some normalization mostly using SELECT DISTINCT on columns and INSERTing results to separate tables. I have added to the old table a column for FK first. Just after the INSERT, I've got its ID and UPDATEed the FK column. When loop finished I've dropped old column leaving only FK column. Similarly with multiple dependent columns. It was much faster than I expected.
I ran (django) python manage.py inspctdb, copied output to models.py and added all those ForeignkeyFields as FKs do not exist on MyISAM. Wrote a little python views.py, urls.py, few templates...TADA
Pandas can return a schema:
pandas.read_csv('data.csv').dtypes
References:
pandas.read_csv
pandas.DataFrame

Categories