You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@arrow.apache.org by "GabsPalomo (via GitHub)" <gi...@apache.org> on 2023/03/15 20:31:02 UTC
[GitHub] [arrow] GabsPalomo opened a new issue, #34578: [R] Error reading .csv.gz files in arrow (straddling object straddles two block boundaries (try to increase block size?)
GabsPalomo opened a new issue, #34578:
URL: https://github.com/apache/arrow/issues/34578
### Describe the usage question you have. Please include as many useful details as possible.
I have a [zip folder with an R project in Google Drive](https://drive.google.com/file/d/1JWIw_j0CUKRe-WV4boCPhsxW6hrR8GRZ/view?usp=share_link) and it contains a script and data folder with 2 .csv.gz files. In the project I have a reproducible example with the code I'm trying to run but getting an error.
Essentially, I have around 300 .csv.gz files organized in different folders and I am trying to use arrow to work with them.
I can open the dataset just fine using open_dataset() but then when I go and collect it I run into this error: ! Invalid: straddling object straddles two block boundaries (try to increase block size?)
Can anyone help me or guide me? I'm new to arrow so it may be that I am defining the schema wrong, or something else. I just don't know how to fix it.
```
library(arrow)
library(dplyr)
# Arrow
data_path <- "./data"
files <- list.files(data_path, full.names = TRUE)
files
length(files)
head(file.info(files), 5)
.schema <- schema(
placekey= utf8(),
parent_placekey=utf8(),
safegraph_brand_ids=utf8(),
location_name= utf8(),
brands=utf8(),
store_id=utf8(),
top_category=utf8(),
sub_category=utf8(),
naics_code= int64(),
latitude= float32(),
longitude= float32(),
street_address=utf8(),
city=utf8(),
region=utf8(),
postal_code=int32(),
open_hours=utf8(),
category_tags=utf8(),
opened_on=date32(),
closed_on=date32(),
tracking_closed_since=date32(),
websites=utf8(),
geometry_type=utf8(),
polygon_wkt=utf8(),
polygon_class=utf8(),
enclosed=bool(),
phone_number=float64(),
is_synthetic=bool(),
includes_parking_lot=bool(),
iso_country_code=utf8(),
wkt_area_sq_meters=uint32(),
date_range_start=timestamp("s"),
date_range_end=timestamp("s"),
raw_visit_counts= int64(),
raw_visitor_counts= int64(),
visits_by_day= utf8(),
poi_cbg=float64(),
visitor_home_cbgs=utf8(),
visitor_home_aggregation=utf8(),
visitor_daytime_cbgs=utf8(),
visitor_country_of_origin=utf8(),
distance_from_home= int64(),
median_dwell=float16(),
bucketed_dwell_times=utf8(),
related_same_day_brand=utf8(),
related_same_month_brand=utf8(),
popularity_by_hour=utf8(),
popularity_by_day=utf8(),
device_type=utf8(),
normalized_visits_by_state_scaling=float32(),
normalized_visits_by_region_naics_visits=float32(),
normalized_visits_by_region_naics_visitors=float32(),
normalized_visits_by_total_visits=float32(),
normalized_visits_by_total_visitors=float32()
)
dat <- open_dataset(data_path, format = 'csv', skip=1, schema = .schema)
dat
dat %>%
select(region, raw_visit_counts) %>%
group_by(region) %>%
summarise(mean_visitors = mean(raw_visit_counts, na.rm=TRUE)) %>%
collect() ->tst
tst
Error in `compute.arrow_dplyr_query()`:
! Invalid: straddling object straddles two block boundaries (try to increase block size?)
Run `rlang::last_error()` to see where the error occurred.
```
### Component(s)
R
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@arrow.apache.org.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [arrow] e-kotov commented on issue #34578: [R] Error reading .csv.gz files in arrow (straddling object straddles two block boundaries (try to increase block size?)
Posted by "e-kotov (via GitHub)" <gi...@apache.org>.
e-kotov commented on issue #34578:
URL: https://github.com/apache/arrow/issues/34578#issuecomment-1470971929
Increase the read block size. The default [seems to be](https://github.com/apache/arrow/blob/6ba2255c83d4813ee75e710d879ef8ff8f7c7539/r/R/csv.R#LL474C33-L474C57) = 1048576L, increasing it by 1000000 to 2048576L seems to work. That, however causes your schema to return a different error, but if you drop the schema, the code below will work. Perhaps you do not need the schema at all or will have to spend some time reviewing it.
```{r}
dat <- open_dataset(data_path, format = 'csv', skip=1, read_options = list(block_size = 2048576L))
dat %>% glimpse()
```
The root of the problem seems to be the extra long strings in the dataset... As noted here https://github.com/huggingface/datasets/issues/836#issuecomment-725552966
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@arrow.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [arrow] thisisnic commented on issue #34578: [R] Error reading .csv.gz files in arrow (straddling object straddles two block boundaries (try to increase block size?)
Posted by "thisisnic (via GitHub)" <gi...@apache.org>.
thisisnic commented on issue #34578:
URL: https://github.com/apache/arrow/issues/34578#issuecomment-1479468765
@GabsPalomo Looks like this is resolved and the example dataset included has now been deleted; I'm going to close this ticket, but let us know if you have any more issues with this.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@arrow.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [arrow] thisisnic closed issue #34578: [R] Error reading .csv.gz files in arrow (straddling object straddles two block boundaries (try to increase block size?)
Posted by "thisisnic (via GitHub)" <gi...@apache.org>.
thisisnic closed issue #34578: [R] Error reading .csv.gz files in arrow (straddling object straddles two block boundaries (try to increase block size?)
URL: https://github.com/apache/arrow/issues/34578
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@arrow.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org